Operasi
mhlo.abs
(mhlo::AbsOp)
Operasi perut
Sintaksis:
operation ::= `mhlo.abs` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Melakukan operasi abs berdasarkan elemen pada tensor operand
dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#abs
Contoh:
%result = mhlo.abs %operand : tensor<3xi32>
Ciri-ciri: AlwaysSpeculatableImplTrait
, Elementwise
, SameOperandsAndResultShape
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
operand | tensor peringkat bilangan bulat tak bertanda 4/8/16/32/64-bit atau tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe bfloat16 atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau integer bertanda terkuantisasi seragam 4/8/16/32-bit atau integer bertanda terkuantisasi seragam 4/8/16/32-bit per sumbu integer bertanda atau 4/8/16/ Seragam 32-bit bilangan bulat tak bertanda terkuantisasi atau seragam 4/8/16/32-bit terkuantisasi per nilai bilangan bulat tak bertanda sumbu |
Hasil:
Hasil | Keterangan |
---|---|
result | tensor peringkat bilangan bulat tak bertanda 4/8/16/32/64-bit atau tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe bfloat16 atau Seragam 4/8/16/32-bit bilangan bulat bertanda terkuantisasi atau seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau 4/8/16/32-bit seragam bilangan bulat tak bertanda terkuantisasi atau 4/8/16/ Seragam 32-bit terkuantisasi per sumbu nilai bilangan bulat tak bertanda tangan |
mhlo.add
(mhlo::TambahOp)
Tambahkan operasi
Sintaksis:
operation ::= `mhlo.add` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Melakukan penambahan dua tensor lhs
dan rhs
berdasarkan elemen dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#add
Contoh:
%result = mhlo.add %lhs, %rhs : tensor<2x2xi32>
Sifat: AlwaysSpeculatableImplTrait
, Commutative
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
lhs | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit bilangan bulat bertanda atau bilangan bulat tak bertanda seragam 4/8/16/32-bit atau bilangan bulat tak bertanda seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau seragam 4/8/16/32-bit terkuantisasi per sumbu nilai bilangan bulat tak bertanda |
rhs | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit bilangan bulat bertanda atau bilangan bulat tak bertanda seragam 4/8/16/32-bit atau bilangan bulat tak bertanda seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau seragam 4/8/16/32-bit terkuantisasi per sumbu nilai bilangan bulat tak bertanda |
Hasil:
Hasil | Keterangan |
---|---|
result | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit bilangan bulat bertanda atau bilangan bulat tak bertanda seragam 4/8/16/32-bit atau bilangan bulat tak bertanda seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau seragam 4/8/16/32-bit terkuantisasi per sumbu nilai bilangan bulat tak bertanda |
mhlo.add_dependency
(mhlo::AddDependencyOp)
Operasi Tambahkan Ketergantungan
Sintaksis:
operation ::= `mhlo.add_dependency` operands attr-dict `:` functional-type(operands, results)
Operasi ini bersifat pribadi untuk kompiler XLA, sehingga belum memiliki spesifikasi.
Secara informal, operasi ini memiliki dua operan: operan data dan token. Output dari operasi ini adalah operan data. Saat digunakan dengan AfterAll, operasi ini memungkinkan pengurutan operasi yang tidak menimbulkan efek samping (operasi yang tidak menghasilkan nilai token).
Contoh:
%1 = mhlo.add_dependency %arg0, %0 : (tensor<3x4xf32>, !mhlo.token) -> tensor<3x4xf32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
operand | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit bilangan bulat bertanda atau bilangan bulat tak bertanda seragam 4/8/16/32-bit terkuantisasi atau tensor peringkat seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau seragam 4/8/16/32-bit terkuantisasi per sumbu nilai integer atau token yang tidak ditandatangani |
token | token |
Hasil:
Hasil | Keterangan |
---|---|
output | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit bilangan bulat bertanda atau bilangan bulat tak bertanda seragam 4/8/16/32-bit terkuantisasi atau tensor peringkat seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau seragam 4/8/16/32-bit terkuantisasi per sumbu nilai integer atau token yang tidak ditandatangani |
mhlo.after_all
(mhlo::AfterAllOp)
Setelah semua operasi
Sintaksis:
operation ::= `mhlo.after_all` $inputs attr-dict
`:` custom<VariadicSameOperandsAndResultType>(ref($inputs), type($inputs), type($result))
Memastikan bahwa operasi yang menghasilkan inputs
dijalankan sebelum operasi apa pun yang bergantung pada result
.
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#after_all
Contoh:
%result = mhlo.after_all %input0, %input1 : !mhlo.token
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
inputs | variadik token |
Hasil:
Hasil | Keterangan |
---|---|
result | token |
mhlo.all_gather
(mhlo::AllGatherOp)
Operasi Semua Kumpulkan
Dalam setiap grup proses di kisi proses, gabungkan nilai tensor operan dari setiap proses sepanjang all_gather_dim
dan menghasilkan tensor hasil. computation
diterapkan secara terpisah untuk setiap operan dalam operands
, menghasilkan satu hasil per operan.
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#all_gather
Contoh:
%result = "mhlo.all_gather"(%operand) {
all_gather_dim = 1 : i64,
replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
// channel_id = 0
channel_handle = #mhlo.channel_handle<handle = 0, type = 0>,
// use_global_device_ids = false
} : (tensor<2x2xf32>) -> tensor<2x4xf32>
Sifat: SameOperandsAndResultElementType
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
all_gather_dim | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
replica_groups | ::mlir::DenseIntElementsAttr | Atribut elemen integer tanpa tanda 64-bit |
channel_handle | ::mlir::mhlo::ChannelHandleAttr | dua bilangan bulat 64-bit 'pegangan' dan 'ketik' |
use_global_device_ids | ::mlir::UnitAttr | atribut satuan |
Operan:
Operan | Keterangan |
---|---|
operands | variadik tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4 /8/16/32/64-bit integer tanpa tanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit atau bilangan bulat tak bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat bertanda atau bilangan bulat bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat tak bertanda nilai-nilai |
Hasil:
Hasil | Keterangan |
---|---|
«tidak disebutkan namanya» | variadik tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4 /8/16/32/64-bit integer tanpa tanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit atau bilangan bulat tak bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat bertanda atau bilangan bulat bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat tak bertanda nilai-nilai |
mhlo.all_reduce
(mhlo::AllReduceOp)
Operasi AllReduce
Dalam setiap grup proses di grid proses, terapkan computation
fungsi reduksi pada nilai tensor operan dari setiap proses dan menghasilkan tensor hasil. computation
diterapkan secara terpisah untuk setiap operan dalam operands
, menghasilkan satu hasil per operan.
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#all_reduce
Contoh:
%result = "mhlo.all_reduce"(%operand) ({
^bb0(%arg0: tensor<f32>, %arg1: tensor<f32>):
%0 = mhlo.add %arg1, %arg2 : tensor<f32>
mhlo.return %0 : tensor<f32>
}) {
replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
// channel_id = 0
channel_handle = #mhlo.channel_handle<handle = 0, type = 0>
// use_global_device_ids = false
} : (tensor<4xf32>) -> tensor<4xf32>
Sifat: InferTensorType
, SingleBlockImplicitTerminator<ReturnOp>
, SingleBlock
Antarmuka: InferShapedTypeOpInterface
, InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
replica_groups | ::mlir::DenseIntElementsAttr | Atribut elemen integer tanpa tanda 64-bit |
channel_handle | ::mlir::mhlo::ChannelHandleAttr | dua bilangan bulat 64-bit 'pegangan' dan 'ketik' |
use_global_device_ids | ::mlir::UnitAttr | atribut satuan |
Operan:
Operan | Keterangan |
---|---|
operands | variadik tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4 /8/16/32/64-bit integer tanpa tanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit atau bilangan bulat tak bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat bertanda atau bilangan bulat bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat tak bertanda nilai-nilai |
Hasil:
Hasil | Keterangan |
---|---|
«tidak disebutkan namanya» | variadik tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4 /8/16/32/64-bit integer tanpa tanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit atau bilangan bulat tak bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat bertanda atau bilangan bulat bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat tak bertanda nilai-nilai |
mhlo.all_to_all
(mhlo::AllToAllOp)
Operasi Semua Ke Semua
Dalam setiap grup proses di kisi proses, pisahkan nilai tensor operand
sepanjang split_dimension
menjadi beberapa bagian, sebarkan bagian yang terpisah di antara proses, gabungkan bagian yang tersebar di sepanjang concat_dimension
dan hasilkan tensor result
.
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#all_to_all
Contoh:
%result = "mhlo.all_to_all"(%operand) {
split_dimension = 1 : i64,
concat_dimension = 0 : i64,
split_count = 2 : i64,
replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
} : (tensor<2x4xf32>) -> tensor<4x2xf32>
Sifat: AlwaysSpeculatableImplTrait
, InferTensorType
, SameOperandsElementType
, SameOperandsShape
, SameVariadicOperandSize
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
split_dimension | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
concat_dimension | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
split_count | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
replica_groups | ::mlir::DenseIntElementsAttr | Atribut elemen integer tanpa tanda 64-bit |
channel_handle | ::mlir::mhlo::ChannelHandleAttr | dua bilangan bulat 64-bit 'pegangan' dan 'ketik' |
Operan:
Operan | Keterangan |
---|---|
operand | variadik tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4 /8/16/32/64-bit integer tanpa tanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit atau bilangan bulat tak bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat bertanda atau bilangan bulat bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat tak bertanda nilai-nilai |
Hasil:
Hasil | Keterangan |
---|---|
«tidak disebutkan namanya» | variadik tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4 /8/16/32/64-bit integer tanpa tanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit atau bilangan bulat tak bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat bertanda atau bilangan bulat bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat tak bertanda nilai-nilai |
mhlo.and
(mhlo::AndOp)
Dan operasi
Sintaksis:
operation ::= `mhlo.and` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Melakukan elemen AND dari dua tensor lhs
dan rhs
dan menghasilkan tensor result
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#and
Contoh:
%result = mhlo.and %lhs, %rhs : tensor<2x2xi32>
Sifat: AlwaysSpeculatableImplTrait
, Commutative
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
lhs | tensor peringkat pred (AKA boolean atau bilangan bulat 1-bit) atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau nilai bilangan bulat tak bertanda 4/8/16/32/64-bit |
rhs | tensor peringkat pred (AKA boolean atau bilangan bulat 1-bit) atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau nilai bilangan bulat tak bertanda 4/8/16/32/64-bit |
Hasil:
Hasil | Keterangan |
---|---|
result | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit bilangan bulat bertanda atau bilangan bulat tak bertanda seragam 4/8/16/32-bit atau bilangan bulat tak bertanda seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau seragam 4/8/16/32-bit terkuantisasi per sumbu nilai bilangan bulat tak bertanda |
mhlo.async_done
(mhlo::AsyncDoneOp)
Operasi AsyncDone
Operasi ini bersifat pribadi untuk kompiler XLA, sehingga belum memiliki spesifikasi.
Secara informal, operasi ini memblokir hingga akhir komputasi asinkron. Ini mengembalikan hasil akhir dari perhitungan asinkron.
Lihat dokumentasi AsyncStart untuk informasi selengkapnya.
Antarmuka: InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
called_computation | ::mlir::FlatSymbolRefAttr | atribut referensi simbol datar |
execution_thread | ::mlir::StringAttr | atribut string |
Operan:
Operan | Keterangan |
---|---|
bundle | async_bundle dengan kombinasi tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit ) atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/ Seragam 32-bit bilangan bulat bertanda terkuantisasi atau seragam 4/8/16/32-bit bilangan bulat tak bertanda terkuantisasi atau seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau 4/8/16/32-bit seragam terkuantisasi per sumbu nilai bilangan bulat yang tidak ditandatangani atau nilai token |
Hasil:
Hasil | Keterangan |
---|---|
«tidak disebutkan namanya» | variadik tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4 /8/16/32/64-bit integer tanpa tanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit atau bilangan bulat tak bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat bertanda atau bilangan bulat bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat tak bertanda nilai atau token atau tupel bersarang dengan kombinasi tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau bilangan bulat 1-bit) atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau tipe kompleks dengan elemen float 32-bit atau float 64-bit atau 4 /8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau nilai bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit atau tensor peringkat 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat bertanda atau 4/8 / Seragam 16/32-bit yang dikuantisasi per sumbu nilai integer atau nilai token yang tidak ditandatangani |
mhlo.async_start
(mhlo::AsyncStartOp)
Operasi AsyncMulai
Operasi ini bersifat pribadi untuk kompiler XLA, sehingga belum memiliki spesifikasi.
Secara informal, operasi ini memulai komputasi asinkron.
Ini digunakan ketika ada fungsi yang berisi menunggu asinkron (seperti DMA) dan komputasi on-thread. Misalnya, suatu fungsi mungkin terdiri dari komputasi, DMA, komputasi lain, DMA kedua, dan komputasi akhir. Ini akan direpresentasikan sebagai async_start diikuti oleh dan async_update dan async_done. async_start akan melakukan komputasi pertama pada thread dan kemudian memulai DMA. async_update akan menunggu hingga DMA selesai jika belum selesai, kemudian menjalankan komputasi kedua dalam fungsi tersebut, dan memulai DMA kedua. Terakhir, async_done akan menunggu pada DMA terakhir ini, lalu menjalankan komputasi terakhir yang perlu dijalankan pada thread dan mengembalikan hasil komputasi akhir tersebut.
operands
diteruskan ke komputasi secara langsung called_computation
adalah fungsi yang akan dijalankan secara asinkron execution_thread
adalah nama thread yang akan menjalankannya. Thread utama disebut "utama". Semua thread mempunyai nama.
Ini mengembalikan semua status yang diperlukan antara operasi asinkron. Setelah penetapan buffer, nilai yang dikembalikan mewakili ruang yang diperlukan untuk menampung input, hasil, dan semua scratchpad yang diperlukan atau diedit oleh operasi asinkron.
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
called_computation | ::mlir::FlatSymbolRefAttr | atribut referensi simbol datar |
execution_thread | ::mlir::StringAttr | atribut string |
Operan:
Operan | Keterangan |
---|---|
inputs | variadik tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4 /8/16/32/64-bit integer tanpa tanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit atau bilangan bulat tak bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat bertanda atau bilangan bulat bertanda 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat tak bertanda nilai atau token atau tupel bersarang dengan kombinasi tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau bilangan bulat 1-bit) atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau tipe kompleks dengan elemen float 32-bit atau float 64-bit atau 4 /8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau nilai bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit atau tensor peringkat 4/8/16/32-bit seragam terkuantisasi per sumbu bilangan bulat bertanda atau 4/8 / Seragam 16/32-bit yang dikuantisasi per sumbu nilai integer atau nilai token yang tidak ditandatangani |
Hasil:
Hasil | Keterangan |
---|---|
«tidak disebutkan namanya» | async_bundle dengan kombinasi tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit ) atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/ Seragam 32-bit bilangan bulat bertanda terkuantisasi atau seragam 4/8/16/32-bit bilangan bulat tak bertanda terkuantisasi atau seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau 4/8/16/32-bit seragam terkuantisasi per sumbu nilai bilangan bulat yang tidak ditandatangani atau nilai token |
mhlo.async_update
(mhlo::AsyncUpdateOp)
Operasi AsyncUpdate
Operasi ini bersifat pribadi untuk kompiler XLA, sehingga belum memiliki spesifikasi.
Secara informal, operasi ini memblokir komputasi asinkron hingga hambatan sinkronisasi. Ini mengembalikan bundle
setelah mengoperasikannya.
Lihat dokumentasi AsyncStart untuk informasi selengkapnya.
Antarmuka: InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
called_computation | ::mlir::FlatSymbolRefAttr | atribut referensi simbol datar |
execution_thread | ::mlir::StringAttr | atribut string |
Operan:
Operan | Keterangan |
---|---|
bundle | async_bundle dengan kombinasi tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit ) atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/ Seragam 32-bit bilangan bulat bertanda terkuantisasi atau seragam 4/8/16/32-bit bilangan bulat tak bertanda terkuantisasi atau seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau 4/8/16/32-bit seragam terkuantisasi per sumbu nilai bilangan bulat yang tidak ditandatangani atau nilai token |
Hasil:
Hasil | Keterangan |
---|---|
«tidak disebutkan namanya» | async_bundle dengan kombinasi tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit ) atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau bilangan bulat tak bertanda 4/8/16/32/64-bit atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau 4/8/16/ Seragam 32-bit bilangan bulat bertanda terkuantisasi atau seragam 4/8/16/32-bit bilangan bulat tak bertanda terkuantisasi atau seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau 4/8/16/32-bit seragam terkuantisasi per sumbu nilai bilangan bulat yang tidak ditandatangani atau nilai token |
mhlo.atan2
(mhlo::Atan2Op)
Operasi Atan2
Sintaksis:
operation ::= `mhlo.atan2` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Melakukan operasi atan2 berdasarkan elemen pada tensor lhs
dan rhs
dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#atan2
Contoh:
%result = mhlo.atan2 %lhs, %rhs : tensor<3xf32>
Sifat: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
lhs | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe bfloat16 atau tipe kompleks dengan elemen float 32-bit atau float 64-bit atau 4/8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau nilai bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit |
rhs | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe bfloat16 atau tipe kompleks dengan elemen float 32-bit atau float 64-bit atau 4/8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau nilai bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit |
Hasil:
Hasil | Keterangan |
---|---|
result | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe bfloat16 atau tipe kompleks dengan elemen float 32-bit atau float 64-bit atau 4/8/16/32-bit bilangan bulat bertanda terkuantisasi seragam atau nilai bilangan bulat tak bertanda terkuantisasi seragam 4/8/16/32-bit |
mhlo.batch_norm_grad
(mhlo::BatchNormGradOp)
Operasi BatchNormGrad
Menghitung gradien beberapa input backpropagating BatchNormTrainingOp dari grad_output
, dan menghasilkan tensor grad_operand
, grad_scale
dan grad_offset
.
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#batch_norm_grad
Contoh:
%grad_operand, %grad_scale, %grad_offset =
"mhlo.batch_norm_grad"(%operand, %scale, %mean, %variance, %grad_output) {
epsilon = 0.0 : f32,
feature_index = 2 : i64
} : (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>,
tensor<2x2x2xf32>) -> (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>)
Sifat: AlwaysSpeculatableImplTrait
, InferTensorType
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
epsilon | ::mlir::FloatAttr | Atribut float 32-bit |
feature_index | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
Operan:
Operan | Keterangan |
---|---|
operand | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
scale | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
mean | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
variance | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
grad_output | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
Hasil:
Hasil | Keterangan |
---|---|
grad_operand | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
grad_scale | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
grad_offset | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
mhlo.batch_norm_inference
(mhlo::BatchNormInferenceOp)
Operasi BatchNormInferensi
Menormalkan tensor operand
di semua dimensi kecuali dimensi feature_index
dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#batch_norm_inference
Contoh:
%result = "mhlo.batch_norm_inference"(%operand, %scale, %offset, %mean, %variance) {
epsilon = 0.0 : f32,
feature_index = 2 : i64
} : (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>) -> tensor<2x2x2xf32>
Sifat: AlwaysSpeculatableImplTrait
, InferTensorType
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
epsilon | ::mlir::FloatAttr | Atribut float 32-bit |
feature_index | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
Operan:
Operan | Keterangan |
---|---|
operand | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
scale | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
offset | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
mean | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
variance | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
Hasil:
Hasil | Keterangan |
---|---|
result | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
mhlo.batch_norm_training
(mhlo::BatchNormTrainingOp)
Operasi BatchNormTraining
Menghitung mean dan varians di seluruh dimensi batch dan spasial serta menormalkan tensor operand
, untuk setiap fitur dalam dimensi feature_index
dan menghasilkan tensor output
, batch_mean
, dan batch_var
.
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#batch_norm_training
Contoh:
%output, %batch_mean, %batch_var = "mhlo.batch_norm_training"(%operand, %scale, %offset) {
epsilon = 0.0 : f32,
feature_index = 2 : i64
} : (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>) -> (tensor<2x2x2xf32>, tensor<2xf32>, tensor<2xf32>)
Sifat: AlwaysSpeculatableImplTrait
, InferTensorType
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
epsilon | ::mlir::FloatAttr | Atribut float 32-bit |
feature_index | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
Operan:
Operan | Keterangan |
---|---|
operand | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
scale | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
offset | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
Hasil:
Hasil | Keterangan |
---|---|
output | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
batch_mean | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
batch_var | Tensor 1D tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau nilai tipe float 16-bit atau float 32-bit atau tipe float 64-bit atau bfloat16 |
mhlo.bitcast
(mhlo::BitcastOp)
Operasi Bitcast
Sintaksis:
operation ::= `mhlo.bitcast` operands attr-dict `:` functional-type(operands, results)
Operasi ini bersifat pribadi untuk kompiler XLA, sehingga belum memiliki spesifikasi.
Secara informal, operasi ini mengubah bentuk masukan sedemikian rupa sehingga susunan fisik elemen tidak berubah.
Operasi ini memerlukan informasi tata letak untuk memahami "susunan fisik elemen", dan dukungan tata letak di MHLO saat ini sedang dalam proses.
Contoh:
%0 = mhlo.bitcast %arg0 : (tensor<3x4xf32>) -> tensor<3x4x1xf32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
operand | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit bilangan bulat bertanda atau bilangan bulat tak bertanda seragam 4/8/16/32-bit atau bilangan bulat tak bertanda seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau seragam 4/8/16/32-bit terkuantisasi per sumbu nilai bilangan bulat tak bertanda |
Hasil:
Hasil | Keterangan |
---|---|
«tidak disebutkan namanya» | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit bilangan bulat bertanda atau bilangan bulat tak bertanda seragam 4/8/16/32-bit atau bilangan bulat tak bertanda seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau seragam 4/8/16/32-bit terkuantisasi per sumbu nilai bilangan bulat tak bertanda |
mhlo.bitcast_convert
(mhlo::BitcastConvertOp)
Operasi Konversi Bitcast
Sintaksis:
operation ::= `mhlo.bitcast_convert` operands attr-dict `:` functional-type(operands, results)
Melakukan operasi bitcast pada tensor operand
dan menghasilkan tensor result
yang mana bit seluruh tensor operand
diinterpretasikan ulang menggunakan jenis tensor result
.
Lihat: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#bitcast_convert
Contoh:
%result = mhlo.bitcast_convert %operand : (tensor<2xf32>) -> tensor<2x4xi8>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
operand | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit bilangan bulat bertanda atau bilangan bulat tak bertanda seragam 4/8/16/32-bit atau bilangan bulat tak bertanda seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau seragam 4/8/16/32-bit terkuantisasi per sumbu nilai bilangan bulat tak bertanda |
Hasil:
Hasil | Keterangan |
---|---|
«tidak disebutkan namanya» | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit bilangan bulat bertanda atau bilangan bulat tak bertanda seragam 4/8/16/32-bit atau bilangan bulat tak bertanda seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau seragam 4/8/16/32-bit terkuantisasi per sumbu nilai bilangan bulat tak bertanda |
mhlo.broadcast
(mhlo::BroadcastOp)
Operasi siaran
Operasi ini sedang dalam proses keluar dari StableHLO, sehingga tidak disertakan dalam spesifikasi: https://github.com/openxla/stablehlo/issues/3
Secara informal, operasi ini melakukan hal yang sama seperti Siaran XLA: https://www.tensorflow.org/xla/operation_semantics#broadcast
Contoh:
%result = mhlo.broadcast %operand, sizes = [1, 2] : (tensor<3xi32>) -> tensor<1x2x3xi32>
Sifat: AlwaysSpeculatableImplTrait
, InferTensorType
, SameOperandsAndResultElementType
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
broadcast_sizes | ::mlir::DenseIntElementsAttr | Atribut elemen integer tanpa tanda 64-bit |
Operan:
Operan | Keterangan |
---|---|
operand | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit bilangan bulat bertanda atau bilangan bulat tak bertanda seragam 4/8/16/32-bit atau bilangan bulat tak bertanda seragam 4/8/16/32-bit terkuantisasi per sumbu bilangan bulat bertanda atau seragam 4/8/16/32-bit terkuantisasi per sumbu nilai bilangan bulat tak bertanda |
Hasil:
Hasil | Keterangan |
---|---|
«tidak disebutkan namanya» | tensor peringkat tipe f8E4M3B11FNUZ atau tipe f8E4M3FN atau tipe f8E4M3FNUZ atau tipe f8E5M2 atau tipe f8E5M2FNUZ atau float 16-bit atau float 32-bit atau float 64-bit atau tipe atau pred bfloat16 (AKA boolean atau integer 1-bit) atau 4/8 /16/32/64-bit integer tak bertanda atau 4/8/16/32/64-bit integer tak bertanda atau tipe kompleks dengan elemen float 32-bit atau elemen float 64-bit atau kuantisasi seragam 4/8/16/32-bit Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
mhlo.broadcast_in_dim
(mhlo :: broadcastIndimop)
Operasi BroadcastIndim
Memperluas dimensi dan/atau peringkat tensor input dengan menduplikasi data dalam tensor operand
dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#broadcast_in_dim
Contoh:
%result = mhlo.broadcast_in_dim %operand, dims = [2, 1] : (tensor<1x3xi32>) -> tensor<2x3x2xi32>
Ciri -ciri: AlwaysSpeculatableImplTrait
, HLO_CompatibleOperandsAndResultElementType
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
broadcast_dimensions | :: mlir :: DenseIntelementsattr | Atribut elemen integer 64-bit |
Operan:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
Hasil:
Hasil | Keterangan |
---|---|
«Tidak disebutkan namanya» | statically shaped tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/ 8/16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-Bit atau Elemen Float 64-Bit atau seragam 4/8/16/32-bit Integer yang ditandatangani terkuantisasi atau integer unifisasi unifisasi unifisasi 4/8/16/32-bit atau seragam kuantisasi per sumbu yang ditandatangani integer atau 4/8/16/32-bit yang diukur per sumbu nilai integer yang tidak ditandatangani tanpa tanda |
mhlo.case
(mhlo :: casedop)
Operasi kasus
Menghasilkan output dari mengeksekusi tepat satu function
dari branches
tergantung pada nilai index
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#case
Contoh:
%result0, %result1 = "mhlo.case"(%index) ({
mhlo.return %result_branch0, %result_branch0 : tensor<2xi64>, tensor<2xi64>
}, {
mhlo.return %result_branch1, %result_branch1 : tensor<2xi64>, tensor<2xi64>
}) : (tensor<i32>) -> (tensor<2xi64>, tensor<2xi64>)
Ciri -ciri: RecursiveMemoryEffects
, SingleBlockImplicitTerminator<ReturnOp>
, SingleBlock
Antarmuka: InferTypeOpInterface
Operan:
Operan | Keterangan |
---|---|
index | Tensor Nilai Integer Tanda Tanda Tanda Tanda |
Hasil:
Hasil | Keterangan |
---|---|
«Tidak disebutkan namanya» | Variadik dari tensor peringkat f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau tipe f8e4m3fnuz atau f8e5m2 atau tipe f8e5m2fnuz atau float 16-bit atau 32-bit atau 64-bit float atau bfloat16 tipe atau pred (aka bool atau 64-bit atau 64-bit float atau bfloat16 atau pred (pari /8/16/32/64-bit Integer SIGLESS atau 4/8/16/32/64-bit Integer Unsigned atau tipe kompleks dengan float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit Integer yang ditandatangani dengan kuantisasi seragam atau nilai integer unigned seragam 4/8/16/32-bit kuantisasi kuantisasi atau tensor peringkat dari 4/8/16/32-bit seragam kuantisasi per sumbu yang ditandatangani integer atau seragam 4/8/8/6/32-bit dikuantisasi seragam kuantisasi per sumbu nilai integer yang tidak ditandatangani atau token |
mhlo.cbrt
(mhlo :: cbrtop)
Operasi CBRT
Sintaksis:
operation ::= `mhlo.cbrt` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Melakukan operasi akar kubik elemen pada tensor operand
dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#cbrt
Contoh:
%result = mhlo.cbrt %operand : tensor<4xf32>
Ciri -ciri: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
operand | Tensor peringkat f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau f8e4m3fnuz tipe atau f8e5m2 tipe atau f8e5m2fnuz tipe atau float 16-bit atau 32-bit float atau 64-bit float atau bfloat16 tipe float dengan 32-bit or float 64-bit atau 64-bit or. 4/8/16/32-bit seragam integer terkuantisasi yang ditandatangani atau 4/8/16/16/32-bit Nilai Integer Unigned Unigned Integer |
Hasil:
Hasil | Keterangan |
---|---|
result | Tensor peringkat f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau f8e4m3fnuz tipe atau f8e5m2 tipe atau f8e5m2fnuz tipe atau float 16-bit atau 32-bit float atau 64-bit float atau bfloat16 tipe float dengan 32-bit or float 64-bit atau 64-bit or. 4/8/16/32-bit seragam integer terkuantisasi yang ditandatangani atau 4/8/16/16/32-bit Nilai Integer Unigned Unigned Integer |
mhlo.ceil
(mhlo :: ceilop)
Operasi ceil
Sintaksis:
operation ::= `mhlo.ceil` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Melakukan langit-langit elemen dari tensor operand
dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#ceil
Contoh:
%result = mhlo.ceil %operand : tensor<5xf32>
Ciri -ciri: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
operand | peringkat tensor f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau tipe f8e4m3fnuz atau tipe f8e5m2 atau tipe f8e5m2fnuz atau float 16-bit atau float 32-bit atau 64-bit float atau bfloat16 nilai jenis |
Hasil:
Hasil | Keterangan |
---|---|
result | peringkat tensor f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau tipe f8e4m3fnuz atau tipe f8e5m2 atau tipe f8e5m2fnuz atau float 16-bit atau float 32-bit atau 64-bit float atau bfloat16 nilai jenis |
mhlo.cholesky
(mhlo :: choleskyop)
Operasi Cholesky
Menghitung dekomposisi cholesky dari batch matriks.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#cholesky
Contoh:
%result = mhlo.cholesky %a, lower = true : tensor<3x3xf32>
Ciri -ciri: AlwaysSpeculatableImplTrait
, InferTensorType
, SameOperandsAndResultElementType
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
lower | ::mlir::BoolAttr | atribut bool |
Operan:
Operan | Keterangan |
---|---|
a | Tensor peringkat f8e4m3b11fnuz tipe atau f8e4m3fn tipe atau f8e4m3fnuz tipe atau f8e5m2 tipe atau f8e5m2fnuz tipe atau float 16-bit atau 32-bit float atau 64-bit float atau bfloat16 tipe float atau float or float 32-bit or 64-bit atau 32-bit or 4-bit or4-bit ors 64-bit or. |
Hasil:
Hasil | Keterangan |
---|---|
«Tidak disebutkan namanya» | Tensor peringkat f8e4m3b11fnuz tipe atau f8e4m3fn tipe atau f8e4m3fnuz tipe atau f8e5m2 tipe atau f8e5m2fnuz tipe atau float 16-bit atau 32-bit float atau 64-bit float atau bfloat16 tipe float atau float or float 32-bit or 64-bit atau 32-bit or 4-bit or4-bit ors 64-bit or. |
mhlo.clamp
(mhlo :: clampop)
Operasi penjepit
Sintaksis:
operation ::= `mhlo.clamp` $min `,` $operand `,` $max attr-dict
`:` custom<SameOperandsAndResultType>(type($min), type($operand), type($max), type($result))
Klem setiap elemen tensor operand
antara nilai minimum dan maksimum dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#clamp
Contoh:
%result = mhlo.clamp %min, %operand, %max : tensor<3xi32>
Ciri -ciri: AlwaysSpeculatableImplTrait
, HLO_BroadcastingElementwise
, InferTensorType
, SameOperandsAndResultElementType
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
min | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
max | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
mhlo.collective_broadcast
(mhlo :: collectiveBroadcastop)
Operasi CollectiveBroadcast
Dalam setiap kelompok proses dalam kisi proses, kirim nilai tensor operand
dari proses sumber ke proses target dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#collective_broadcast
Contoh:
%result = "mhlo.collective_broadcast"(%operand) {
replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>,
channel_handle = #mhlo.channel_handle<handle = 0, type = 0>
} : (tensor<1x2xi64>) -> tensor<1x2xi64>
Ciri -ciri: CompatibleOperandsAndResultType
Antarmuka: InferShapedTypeOpInterface
, InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
replica_groups | :: mlir :: DenseIntelementsattr | Atribut elemen integer 64-bit |
channel_handle | :: mlir :: mhlo :: channelHandleAttr | dua handle 64-bit integers 'dan' type ' |
Operan:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
Hasil:
Hasil | Keterangan |
---|---|
«Tidak disebutkan namanya» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
mhlo.collective_permute
(mhlo :: collectivePermuteop)
Operasi CollectivePermute
Dalam setiap kelompok proses dalam kisi proses, mengirimkan nilai tensor operand
dari proses sumber ke proses target dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#collective_permute
Contoh:
%result = "mhlo.collective_permute"(%operand) {
source_target_pairs = dense<[[0, 1], [1, 2]]> : tensor<2x2xi64>,
// channel_id = 0
channel_handle = #mhlo.channel_handle<handle = 0, type = 0>
} : (tensor<4x2xf32>) -> tensor<4x2xf32>
Ciri -ciri: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
source_target_pairs | :: mlir :: DenseIntelementsattr | Atribut elemen integer 64-bit |
channel_handle | :: mlir :: mhlo :: channelHandleAttr | dua handle 64-bit integers 'dan' type ' |
Operan:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
Hasil:
Hasil | Keterangan |
---|---|
«Tidak disebutkan namanya» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
mhlo.compare
(mhlo :: compareop)
Bandingkan operasi
Sintaksis:
operation ::= `mhlo.compare` $comparison_direction `,` $lhs `,` $rhs (`,` $compare_type^)?
attr-dict `:` functional-type(operands, results)
Melakukan perbandingan elemen dari tensor lhs
dan rhs
menurut comparison_direction
dan compare_type
, dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#compare
Contoh:
%result = mhlo.compare LT, %lhs, %rhs, FLOAT : (tensor<2xf32>, tensor<2xf32>) -> tensor<2xi1>
Ciri -ciri: AlwaysSpeculatableImplTrait
, Elementwise
, InferTensorType
, SameOperandsAndResultShape
, SameOperandsElementType
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
comparison_direction | :: mlir :: mhlo :: comparisondirectionattr | Operasi perbandingan mana yang harus dilakukan. |
compare_type | :: mlir :: mhlo :: comparisontypeattr | Jenis perbandingan mana yang akan digunakan. |
Operan:
Operan | Keterangan |
---|---|
lhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
rhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
Hasil:
Hasil | Keterangan |
---|---|
«Tidak disebutkan namanya» | peringkat tensor nilai pred (alias boolean atau 1-bit integer)) |
mhlo.complex
(mhlo :: complexop)
Operasi yang kompleks
Sintaksis:
operation ::= `mhlo.complex` operands attr-dict
`:` custom<ComplexOpType>(type($lhs), type($rhs), type($result))
Melakukan konversi elemen-bijaksana ke nilai kompleks dari sepasang nilai nyata dan imajiner, lhs
dan rhs
, dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#complex
Contoh:
%result = mhlo.complex %lhs, %rhs : tensor<2xcomplex<f32>>
Ciri -ciri: AlwaysSpeculatableImplTrait
, Elementwise
, SameOperandsAndResultShape
, SameOperandsElementType
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
lhs | Tensor peringkat float 32-bit atau nilai apung 64-bit |
rhs | Tensor peringkat float 32-bit atau nilai apung 64-bit |
Hasil:
Hasil | Keterangan |
---|---|
result | peringkat tensor tipe kompleks dengan nilai float 32-bit atau 64-bit float elemen |
mhlo.composite
(mhlo :: compositeop)
Operasi gabungan
Sintaksis:
operation ::= `mhlo.composite` $name $inputs attr-dict `:` functional-type(operands, results)
Merangkum operasi yang dibuat (terdiri) dari operasi stableHlo lainnya, mengambil inputs
dan composite_attributes
dan menghasilkan results
. Semantik OP diimplementasikan oleh atribut decomposition
. OP composite
dapat diganti dengan dekomposisi tanpa mengubah semantik program. Dalam kasus di mana inlining dekomposisi tidak memberikan semantik OP yang sama, lebih suka menggunakan custom_call
.
Bidang version
(default ke 0
) digunakan untuk menunjukkan ketika semantik komposit berubah.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#composite
Contoh:
%results = mhlo.composite "my.op" %arg0, %arg1 {
decomposition = @my_op,
composite_attributes = { my_attribute = "my_value" },
version = 1 : i32
} : (tensor<f32>, tensor<f32>) -> tensor<f32>
Antarmuka: SymbolUserOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
name | ::mlir::StringAttr | atribut string |
composite_attributes | ::mlir::DictionaryAttr | kamus nilai atribut bernama |
decomposition | ::mlir::FlatSymbolRefAttr | atribut referensi simbol datar |
version | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 32-bit |
Operan:
Operan | Keterangan |
---|---|
inputs | Variadik dari tensor peringkat f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau tipe f8e4m3fnuz atau f8e5m2 atau tipe f8e5m2fnuz atau float 16-bit atau 32-bit atau 64-bit float atau bfloat16 tipe atau pred (aka bool atau 64-bit atau 64-bit float atau bfloat16 atau pred (pari /8/16/32/64-bit Integer SIGLESS atau 4/8/16/32/64-bit Integer Unsigned atau tipe kompleks dengan float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit seragam integer yang ditandatangani kuantisasi atau seragam unifisasi unit unigned unifisasi atau 4/8/16/16/32-bit kuantisasi per sumbu yang ditandatangani integer atau 4/8/16/32-bit yang dikuantisasi per sumbu Integer Unign Unigned Integer Unigned nilai atau token atau tuple bersarang dengan kombinasi tensor peringkat f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau tipe f8e4m3fnuu atau tipe f8e5m2 atau f8e5m2fnuz tipe orak orka (misal booat atau booat atau 64-bit float atau 64-bit (64-bit float atau 64-bit (64-bit (64-bit oraAn 64M2 atau 1-bit integer) atau 4/8/16/32/64-bit integer tanpa tanda atau 4/8/16/32/64-bit unsigned integer atau tipe kompleks dengan float 32-bit atau elemen float 64-bit atau 4 /8/16/32-bit seragam integer terkuantisasi yang ditandatangani atau 4/8/16/6/32-bit Nilai integer unigned kuantisasi atau tensor peringkat dari seragam 4/8/16/32-bit yang dikuantisasi per sumbu integer atau 4/8 /16/32-bit seragam kuantisasi per sumbu nilai integer yang tidak ditandatangani atau nilai token |
Hasil:
Hasil | Keterangan |
---|---|
«Tidak disebutkan namanya» | Variadik dari tensor peringkat f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau tipe f8e4m3fnuz atau f8e5m2 atau tipe f8e5m2fnuz atau float 16-bit atau 32-bit atau 64-bit float atau bfloat16 tipe atau pred (aka bool atau 64-bit atau 64-bit float atau bfloat16 atau pred (pari /8/16/32/64-bit Integer SIGLESS atau 4/8/16/32/64-bit Integer Unsigned atau tipe kompleks dengan float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit seragam integer yang ditandatangani kuantisasi atau seragam unifisasi unit unigned unifisasi atau 4/8/16/16/32-bit kuantisasi per sumbu yang ditandatangani integer atau 4/8/16/32-bit yang dikuantisasi per sumbu Integer Unign Unigned Integer Unigned nilai atau token atau tuple bersarang dengan kombinasi tensor peringkat f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau tipe f8e4m3fnuu atau tipe f8e5m2 atau f8e5m2fnuz tipe orak orka (misal booat atau booat atau 64-bit float atau 64-bit (64-bit float atau 64-bit (64-bit (64-bit oraAn 64M2 atau 1-bit integer) atau 4/8/16/32/64-bit integer tanpa tanda atau 4/8/16/32/64-bit unsigned integer atau tipe kompleks dengan float 32-bit atau elemen float 64-bit atau 4 /8/16/32-bit seragam integer terkuantisasi yang ditandatangani atau 4/8/16/6/32-bit Nilai integer unigned kuantisasi atau tensor peringkat dari seragam 4/8/16/32-bit yang dikuantisasi per sumbu integer atau 4/8 /16/32-bit seragam kuantisasi per sumbu nilai integer yang tidak ditandatangani atau nilai token |
mhlo.compute_reshape_shape
(mhlo :: computereshapeshapeop)
Operasi ComputereShapeshape
Sintaksis:
operation ::= `mhlo.compute_reshape_shape` operands attr-dict `:` functional-type(operands, results)
Operasi ini sedang dalam proses, jadi belum termasuk dalam spesifikasi: https://github.com/openxla/stableHlo/issues/8
Secara informal, operasi ini menghitung output_shape untuk Dynamicreshapeop dari jumlah elemen num_elements
dalam operan Dynamicreshapeop dan bentuk dynamic_shape
yang disediakan untuk pembentukan kembali TF: https://www.tensorflow.org/api_docs/python/treshape
Misalnya, untuk num_elements = 12
dan dynamic_shape = [2, -1]
, result
akan menjadi [2, 6]
. Jika operan tidak valid (misalnya, dimensi tidak secara merata membagi jumlah elemen, atau jika ada beberapa nilai -1 dalam dimensi), ini mengarah pada perilaku yang tidak ditentukan.
Contoh:
%result = mhlo.compute_reshape_shape %num_elements, %dynamic_shape
: (index, tensor<2xi32>) -> tensor<2xi32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
num_elements | indeks |
dynamic_shape | 1D tensor integer atau nilai indeks |
Hasil:
Hasil | Keterangan |
---|---|
result | 1D tensor integer atau nilai indeks |
mhlo.concatenate
(mhlo :: concatenateop)
Operasi gabungan
Menggabungkan sejumlah variad tensor dalam inputs
sepanjang dimensi dimension
dalam urutan yang sama dengan argumen yang diberikan dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#concateNate
Contoh:
%result = mhlo.concatenate %input0, %input1, dim = 0 : (tensor<3x2xi64>, tensor<1x2xi64>) -> tensor<4x2xi64>
Ciri-ciri: AlwaysSpeculatableImplTrait
, SameOperandsAndResultElementType
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
dimension | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
Operan:
Operan | Keterangan |
---|---|
val | Variadik dari tensor peringkat f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau tipe f8e4m3fnuz atau f8e5m2 atau tipe f8e5m2fnuz atau float 16-bit atau 32-bit atau 64-bit float atau bfloat16 tipe atau pred (aka bool atau 64-bit atau 64-bit float atau bfloat16 atau pred (pari /8/16/32/64-bit Integer SIGLESS atau 4/8/16/32/64-bit Integer Unsigned atau tipe kompleks dengan float 32-bit atau elemen float 64-bit atau 4/8/16/32-bit seragam integer yang ditandatangani kuantisasi atau seragam unifisasi unit unigned unifisasi atau 4/8/16/16/32-bit kuantisasi per sumbu yang ditandatangani integer atau 4/8/16/32-bit yang dikuantisasi per sumbu Integer Unign Unigned Integer Unigned nilai-nilai |
Hasil:
Hasil | Keterangan |
---|---|
«Tidak disebutkan namanya» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
mhlo.constant
(mhlo :: constantop)
Operasi Konstan
Menghasilkan tensor output
dari value
konstan.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#constant
Contoh:
%output = mhlo.constant dense<[[0.0, 1.0], [2.0, 3.0]]> : tensor<2x2xf32>
Ciri-ciri: AlwaysSpeculatableImplTrait
, ConstantLike
Antarmuka: ConditionallySpeculatable
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
value | :: mlir :: elementsattr | atribut vektor/tensor konstan |
Hasil:
Hasil | Keterangan |
---|---|
output | statically shaped tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/ 8/16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-Bit atau Elemen Float 64-Bit atau seragam 4/8/16/32-bit Integer yang ditandatangani terkuantisasi atau integer unifisasi unifisasi unifisasi 4/8/16/32-bit atau seragam kuantisasi per sumbu yang ditandatangani integer atau 4/8/16/32-bit yang diukur per sumbu nilai integer yang tidak ditandatangani tanpa tanda |
mhlo.convert
(mhlo :: convertop)
Konversi operasi
Sintaksis:
operation ::= `mhlo.convert` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Melakukan konversi elemen-bijaksana dari satu jenis elemen ke jenis lain pada tensor operand
dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#convert
Contoh:
%result = mhlo.convert %operand : (tensor<3xi32>) -> tensor<3xcomplex<f32>>
Ciri -ciri: AlwaysSpeculatableImplTrait
, Elementwise
, SameOperandsAndResultShape
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
mhlo.convolution
(mhlo :: convolutionop)
Operasi konvolusi
Sintaksis:
operation ::= `mhlo.convolution` `(`operands`)`
`dim_numbers` `=` custom<ConvolutionDimensions>($dimension_numbers) `,`
`window` `=` `{` custom<WindowAttributes>($window_strides, $padding,
$lhs_dilation, $rhs_dilation,
$window_reversal) `}`
attr-dict `:` functional-type(operands, results)
Menghitung produk titik antara jendela lhs
dan irisan rhs
dan menghasilkan result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#convolution
Contoh:
%result = "mhlo.convolution"(%lhs, %rhs) {
window_strides = dense<4> : tensor<2xi64>,
padding = dense<0> : tensor<2x2xi64>,
lhs_dilation = dense<2> : tensor<2xi64>,
rhs_dilation = dense<1> : tensor<2xi64>,
window_reversal = dense<false> : tensor<2xi1>,
dimension_numbers = #mhlo.conv<[b, 0, 1, f]x[0, 1, i, o]->[b, 0, 1, f]>,
feature_group_count = 1 : i64,
batch_group_count = 1 : i64,
precision_config = [#stablehlo<precision DEFAULT>, #stablehlo<precision DEFAULT>]
} : (tensor<1x4x4x1xi32>, tensor<3x3x1x1xi32>) -> tensor<1x2x2x1xi32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
window_strides | :: mlir :: DenseIntelementsattr | Atribut elemen integer 64-bit |
padding | :: mlir :: DenseIntelementsattr | Atribut elemen integer 64-bit |
lhs_dilation | :: mlir :: DenseIntelementsattr | Atribut elemen integer 64-bit |
rhs_dilation | :: mlir :: DenseIntelementsattr | Atribut elemen integer 64-bit |
window_reversal | :: mlir :: DenseElementsattr | atribut vektor/tensor boolean konstan |
dimension_numbers | :: mlir :: mhlo :: convdimensionnumbersattr | Struktur Informasi Dimensi untuk Konvensi |
feature_group_count | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
batch_group_count | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
precision_config | ::mlir::ArrayAttr | Atribut Konfigurasi Presisi |
Operan:
Operan | Keterangan |
---|---|
lhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
rhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
Hasil:
Hasil | Keterangan |
---|---|
«Tidak disebutkan namanya» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang dikuantisasi per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani 4/8/16/32 |
mhlo.copy
(mhlo :: copyop)
Operasi Salin
Sintaksis:
operation ::= `mhlo.copy` operands attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Operasi ini bersifat pribadi untuk kompiler XLA, sehingga belum memiliki spesifikasi.
Secara informal, operasi ini salinan operand
. Bergantung pada metadata yang terpasang pada operasi, itu bisa berperilaku sangat berbeda dari no-op.
Contoh:
%0 = mhlo.copy %arg0 : tensor<f32>
Ciri -ciri: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
Elementwise
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
cross_program_prefetch_index | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 32-bit |
Operan:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang diukur per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani atau 4/6/32 Tuple token atau bersarang dengan kombinasi tensor peringkat dari jenis f8e4m3b11fnuz atau tipe f8e4m3fn atau tipe f8e4m3fnuz atau tipe f8e5m2 atau jenis f8e5m2fnuz atau float 16-bit atau float 32-bit atau 64-bit atau bflin atau bflfnuz atau 16-bit float atau 32-bit float atau 64-bit atau bflfnuz atau bflfnuz atau 16-bit float atau 32-bit float atau 64-bit atau bflfnuz atau bflfnuz atau 16-bit float atau 32-bit float atau 64-bit atau bflfloat16 -bit integer) atau 4/8/16/32/64-bit integer suthless atau 4/8/16/32/64-bit Integer atau tipe kompleks tanpa unggulan dengan float 32-bit atau elemen float 64-bit atau 4/8 /16/32-bit seragam integer terkuantisasi yang ditandatangani atau 4/8/16/32-bit seragam nilai integer unigned kuantisasi atau tensor peringkat 4/8/16/32-bit seragam kuantisasi per integer yang ditandatangani atau 4/8/16 /32-bit seragam kuantisasi per sumbu nilai integer yang tidak ditandatangani atau nilai token |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8 /16/32/64-bit Integer Tanda Tanda Tanda atau 4/8/16/32/64-bit Integer Unsigned atau Tipe Kompleks dengan Float 32-bit atau elemen float 64-bit atau seragam 4/8/16/32-bit yang dikuantisasi dikuantisasi terkuantisasi Integer yang ditandatangani atau seragam unifisasi unifisasi unifisasi unifisasi unifisasi unifisasi atau 4/8/16/32-bit yang diukur per sumbu yang ditandatangani integer atau seragam integer yang tidak ditandatangani atau 4/6/32 Tuple token atau bersarang dengan kombinasi tensor peringkat dari jenis f8e4m3b11fnuz atau tipe f8e4m3fn atau tipe f8e4m3fnuz atau tipe f8e5m2 atau jenis f8e5m2fnuz atau float 16-bit atau float 32-bit atau 64-bit atau bflin atau bflfnuz atau 16-bit float atau 32-bit float atau 64-bit atau bflfnuz atau bflfnuz atau 16-bit float atau 32-bit float atau 64-bit atau bflfnuz atau bflfnuz atau 16-bit float atau 32-bit float atau 64-bit atau bflfloat16 -bit integer) atau 4/8/16/32/64-bit integer suthless atau 4/8/16/32/64-bit Integer atau tipe kompleks tanpa unggulan dengan float 32-bit atau elemen float 64-bit atau 4/8 /16/32-bit seragam integer terkuantisasi yang ditandatangani atau 4/8/16/32-bit seragam nilai integer unigned kuantisasi atau tensor peringkat 4/8/16/32-bit seragam kuantisasi per integer yang ditandatangani atau 4/8/16 /32-bit seragam kuantisasi per sumbu nilai integer yang tidak ditandatangani atau nilai token |
mhlo.cosine
(mhlo :: cosineop)
Operasi kosinus
Sintaksis:
operation ::= `mhlo.cosine` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Melakukan operasi kosinus elemen pada tensor operand
dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#cosine
Contoh:
%result = mhlo.cosine %operand : tensor<2xf32>
Ciri -ciri: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
operand | Tensor peringkat f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau f8e4m3fnuz tipe atau f8e5m2 tipe atau f8e5m2fnuz tipe atau float 16-bit atau 32-bit float atau 64-bit float atau bfloat16 tipe float dengan 32-bit or float 64-bit atau 64-bit or. 4/8/16/32-bit seragam integer terkuantisasi yang ditandatangani atau 4/8/16/16/32-bit Nilai Integer Unigned Unigned Integer |
Hasil:
Hasil | Keterangan |
---|---|
result | Tensor peringkat f8e4m3b11fnuz tipe atau tipe f8e4m3fn atau f8e4m3fnuz tipe atau f8e5m2 tipe atau f8e5m2fnuz tipe atau float 16-bit atau 32-bit float atau 64-bit float atau bfloat16 tipe float dengan 32-bit or float 64-bit atau 64-bit or. 4/8/16/32-bit seragam integer terkuantisasi yang ditandatangani atau 4/8/16/16/32-bit Nilai Integer Unigned Unigned Integer |
mhlo.count_leading_zeros
(mhlo :: clzop)
Operasi CLZ
Sintaksis:
operation ::= `mhlo.count_leading_zeros` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Melakukan penghitungan elemen dari jumlah bit nol terkemuka dalam tensor operand
dan menghasilkan tensor result
.
Lihat: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#count_leading_zeros
Contoh:
%result = mhlo.count_leading_zeros %operand : tensor<2x2xi8>
Ciri -ciri: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operan:
Operan | Keterangan |
---|---|
operand | Tensor peringkat 4/8/16/32/64-bit Integer SIGLESS atau 4/8/16/32/64-bit Nilai Integer Unsigned Integer |
Hasil:
Hasil | Keterangan |
---|---|
result | Tensor peringkat 4/8/16/32/64-bit Integer SIGLESS atau 4/8/16/32/64-bit Nilai Integer Unsigned Integer |
mhlo.create_token
(mhlo :: createTokenop)
Operasi Kreatetoken
Sintaksis:
operation ::= `mhlo.create_token` attr-dict `:` type(results)
Operasi ini sedang dalam perjalanan keluar dari stableHlo, jadi tidak termasuk dalam spesifikasi: https://github.com/openxla/stableHlo/issues/3
Secara informal, operasi ini melakukan hal yang sama seperti afterallop dengan 0 input: https://github.com/openxla/staberhlo/blob/main/docs/spec.md#after_all
Contoh:
%output = mhlo.create_token : !mhlo.token
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Hasil:
Hasil | Keterangan |
---|---|
output | token |
mhlo.cross-replica-sum
(mhlo :: crossreplicasumop)
Operasi CrossReplicasum
Operasi ini sedang dalam perjalanan keluar dari stableHlo, jadi tidak termasuk dalam spesifikasi: https://github.com/openxla/stableHlo/issues/3
Secara informal, operasi ini melakukan hal yang sama dengan AllReduceop dengan channel_id = 0
, use_global_device_ids = false
dan computation
menerapkan penambahan: https://github.com/openxla/stableHlo/blob/main/docs/spec.md#all_reduce
Contoh:
%result = "mhlo.cross-replica-sum"(%operand) {
replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>
} : (tensor<4xf32>) -> tensor<4xf32>
Ciri -ciri: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
Antarmuka: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
replica_groups | :: mlir :: DenseIntelementsattr | Atribut elemen integer 64-bit |
Operan:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.cstr_reshapable
(mhlo::CstrReshapableOp)
CstrReshapable operation
Sintaksis:
operation ::= `mhlo.cstr_reshapable` operands attr-dict `:` functional-type(operands, results)
This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8
Informally, this operation creates a witness on the constraint that ComputeReshapeShape would succeed with the provided operands.
Contoh:
%result = mhlo.cstr_reshapable %num_elements, %dynamic_shape
: (index, tensor<3xi32>) -> !shape.witness
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
num_elements | indeks |
dynamic_shape | 1D tensor of integer or index values |
Hasil:
Hasil | Keterangan |
---|---|
result |
mhlo.custom_call
(mhlo::CustomCallOp)
CustomCall operation
Sintaksis:
operation ::= `mhlo.custom_call` custom<CustomCallTarget>($call_target_name) `(` $inputs `)`
attr-dict `:` functional-type(operands, results)
Encapsulates an implementation-defined operation call_target_name
that takes inputs
and called_computations
and produces results
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#custom_call
Contoh:
%results = "mhlo.custom_call"(%input0) {
call_target_name = "foo",
has_side_effect = false,
backend_config = "bar",
api_version = 1 : i32,
called_computations = [@foo]
} : (tensor<f32>) -> tensor<f32>
A custom call invokes code external to XLA. The `inputs` are passed to the
external code, and the external code is expected to produce a result of the
given type. The exact mechanism is backend-specific. For example, in the CPU
backend, a call instruction is emitted which targets a symbol with the name
`call_target_name`.
If XLA runtime is enabled for a backend, then custom calls use the runtime
custom call calling convention to call into the external functions. This
calling convention defines an ABI for encoding arguments, attributes and
results.
Depending on the API version there are two ways to pass extra bits of static
information to the external function:
1. For `API_VERSION_TYPED_FFI` custom calls `backend_config` must be a
dictionary attribute, that will be encoded according to the custom call
calling convention and passed to the external function as the attributes
argument. External code is expected to use declarative bindings (see
`xla/runtime/custom_call.h`) to decode them at run time. These custom
calls are only supported if XLA uses XLA runtime.
2. For previous API versions it is the user responsibility to encode extra
bits of static information as a string `backend_config` attribute, and
decode it at run time.
Antarmuka: MemoryEffectOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
call_target_name | ::mlir::StringAttr | atribut string |
has_side_effect | ::mlir::BoolAttr | atribut bool |
backend_config | ::mlir::Atribut | string attribute or dictionary of named attribute values |
api_version | ::mlir::mhlo::CustomCallApiVersionAttr | Custom call API version |
called_computations | ::mlir::ArrayAttr | atribut array ref simbol datar |
custom_call_schedule | ::mlir::mhlo::CustomCallScheduleAttr | Specifies the desired schedule for the custom-call. |
operand_layouts | ::mlir::ArrayAttr | Array of layout (1D tensor of index type) attributes |
result_layouts | ::mlir::ArrayAttr | Array of layout (1D tensor of index type) attributes |
output_operand_aliases | ::mlir::ArrayAttr | Aliasing attribute for outputs and operands of CustomCall |
Operands:
Operan | Keterangan |
---|---|
inputs | variadic of tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token or nested tuple with any combination of tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | variadic of tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token or nested tuple with any combination of tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token values |
mhlo.divide
(mhlo::DivOp)
Operasi div
Sintaksis:
operation ::= `mhlo.divide` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise division of dividend lhs
and divisor rhs
tensors and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#divide
Contoh:
%result = mhlo.divide %lhs, %rhs : tensor<4xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.domain
(mhlo::DomainOp)
Domain operation
This operation is private to the XLA compiler, so it is does not yet have a specification.
Informally, these operations are used to group instructions with the same DomainMetadata property. ShardingMetadata is the main use case today to group instructions on the same device. Domain instructions provide two major benefits:
- Prevent unintentionally optimizing instructions across domains.
- Automatically assign the metadata of the instructions created in the domain. Without domain instructions, each HLO optimization pass would have to check and propagate the metadata, which would be easy to miss and also adds complexity to the compiler. Since domain instructions connect two different domains, each domain instruction is associated with two DomainMetadata -- one on the operand side and one on the user side of the domain.
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
kind | ::mlir::mhlo::DomainKindAttr | Kind of domain metatdata attached to an HLO domain. |
entry_metadata | ::mlir::StringAttr | atribut string |
exit_metadata | ::mlir::StringAttr | atribut string |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token |
mhlo.dot
(mhlo::DotOp)
Operasi titik
This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/3
Informally, this operation does the same thing as XLA's Dot: https://www.tensorflow.org/xla/operation_semantics#dot
Contoh:
%0 = mhlo.dot %arg0, %arg1 : (tensor<1x2xi32>, tensor<2x1xi32>) -> tensor<1x1xi32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
precision_config | ::mlir::ArrayAttr | Precision Config attribute |
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.dot_general
(mhlo::DotGeneralOp)
DotGeneral operation
Computes dot products between slices of lhs
and slices of rhs
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#dot_general
Contoh:
%result = "mhlo.dot_general"(%lhs, %rhs) {
dot_dimension_numbers = #mhlo.dot<
lhs_batching_dimensions = [0],
rhs_batching_dimensions = [0],
lhs_contracting_dimensions = [2],
rhs_contracting_dimensions = [1]
>,
precision_config = [#stablehlo<precision DEFAULT>, #stablehlo<precision DEFAULT>]
} : (tensor<2x2x2xi32>, tensor<2x2x2xi32>) -> tensor<2x2x2xi32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
dot_dimension_numbers | ::mlir::mhlo::DotDimensionNumbersAttr | Attribute that models the dimension information for dot. |
precision_config | ::mlir::ArrayAttr | Precision Config attribute |
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.dynamic_broadcast_in_dim
(mhlo::DynamicBroadcastInDimOp)
DynamicBroadcastInDim operation
This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8
Informally, this operation does the same thing as BroadcastInDimOp except that the result shape is specified dynamically via output_dimensions
: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#broadcast_in_dim
It also accepts optional attributes to express static knowledge about the expanding behavior of dimensions. If not specified, all dimensions are assumed to be possibly expanding. The sets of dimensions that are known to be expanding and the set of dimensions that are known to be non-expanding must be disjoint and they must be a subset of the operand's dimensions.
Ciri-ciri: AlwaysSpeculatableImplTrait
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
broadcast_dimensions | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
known_expanding_dimensions | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
known_nonexpanding_dimensions | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
output_dimensions | 1D tensor of index or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.dynamic_conv
(mhlo::DynamicConvOp)
DynamicConv operation
This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8
Informally, this operation does the same thing as ConvolutionOp except that padding
is specified dynamically via d_padding
: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#convolution
Contoh:
%result = "mhlo.dynamic_conv"(%lhs, %rhs, %d_padding) {
window_strides = dense<4> : tensor<2xi64>,
lhs_dilation = dense<2> : tensor<2xi64>,
rhs_dilation = dense<1> : tensor<2xi64>,
window_reversal = dense<false> : tensor<2xi1>,
dimension_numbers = #mhlo.conv<[b, 0, 1, f]x[0, 1, i, o]->[b, 0, 1, f]>,
feature_group_count = 1 : i64,
batch_group_count = 1 : i64,
precision_config = [#stablehlo<precision DEFAULT>, #stablehlo<precision DEFAULT>]
} : (tensor<1x4x4x1xi32>, tensor<3x3x1x1xi32>, tensor<2x2xi64>) -> tensor<1x2x2x1xi32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
window_strides | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
padding | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
lhs_dilation | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
rhs_dilation | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
window_reversal | ::mlir::DenseElementsAttr | constant boolean vector/tensor attribute |
dimension_numbers | ::mlir::mhlo::ConvDimensionNumbersAttr | Structure of dimension information for conv op |
feature_group_count | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
batch_group_count | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
precision_config | ::mlir::ArrayAttr | Precision Config attribute |
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
d_padding | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.dynamic_gather
(mhlo::DynamicGatherOp)
DynamicGather operation
This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8
Informally, this operation does the same thing as GatherOp except that slice_sizes
are specified dynamically: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#gather
Contoh:
%result = "mhlo.dynamic_gather"(%operand, %start_indices, %slice_sizes) {
dimension_numbers = #mhlo.gather<
offset_dims = [2, 3],
collapsed_slice_dims = [0],
start_index_map = [0, 2],
index_vector_dim = 2>,
indices_are_sorted = false
} : (tensor<3x4x2xi32>, tensor<2x3x2xi64>, tensor<3xi64>) -> tensor<2x3x2x2xi32>
Traits: AlwaysSpeculatableImplTrait
, InferTensorType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
dimension_numbers | ::mlir::mhlo::GatherDimensionNumbersAttr | Attribute that models the dimension information for gather |
indices_are_sorted | ::mlir::BoolAttr | atribut bool |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
start_indices | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
slice_sizes | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.dynamic_iota
(mhlo::DynamicIotaOp)
DynamicIota operation
This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8
Informally, this operation does the same thing as IotaOp except that the result shape is specified dynamically via output_shape
: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#iota
Contoh:
%0 = mhlo.dynamic_iota %arg0, dim = 0 : (tensor<1xindex>) -> tensor<4xi32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
iota_dimension | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
Operands:
Operan | Keterangan |
---|---|
output_shape | 1D tensor of index or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.dynamic_pad
(mhlo::DynamicPadOp)
DynamicPad operation
Sintaksis:
operation ::= `mhlo.dynamic_pad` operands attr-dict `:` functional-type(operands, results)
Dynamically Pads the operand
, with amount of padding added at low-end/high-end/interior is passed through input tensors.
Ciri-ciri: AlwaysSpeculatableImplTrait
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
padding_value | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
edge_padding_low | 1D tensor of index or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
edge_padding_high | 1D tensor of index or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
interior_padding | 1D tensor of index or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.dynamic_reshape
(mhlo::DynamicReshapeOp)
DynamicReshape operation
Sintaksis:
operation ::= `mhlo.dynamic_reshape` operands attr-dict `:` functional-type(operands, results)
This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8
Informally, this operation does the same thing as ReshapeOp except that the result shape is specified dynamically via output_shape
: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reshape
Contoh:
%0 = mhlo.dynamic_reshape %arg0, %shape : (tensor<?xf32>, tensor<2xindex>) -> tensor<?x?xf32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
output_shape | 1D tensor of index or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.dynamic_slice
(mhlo::DynamicSliceOp)
DynamicSlice operation
Extracts a slice from the operand
using dynamically-computed starting indices and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#dynamic_slice
Contoh:
%result = mhlo.dynamic_slice %operand, %start_indices0, %start_indices1, sizes = [2, 2]
: (tensor<4x4xi32>, tensor<i64>, tensor<i64>) -> tensor<2x2xi32>
Traits: AlwaysSpeculatableImplTrait
, InferTensorType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
slice_sizes | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
start_indices | variadic of 0D tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.dynamic_update_slice
(mhlo::DynamicUpdateSliceOp)
DynamicUpdateSlice operation
Sintaksis:
operation ::= `mhlo.dynamic_update_slice` operands attr-dict `:` functional-type(operands, results)
Produces a result
tensor which is equal to the operand
tensor except that the slice starting at start_indices
is updated with the values in update
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#dynamic_update_slice
Contoh:
%result = mhlo.dynamic_update_slice %operand, %update, %start_indices0, %start_indices1
: (tensor<4x4xi32>, tensor<2x2xi32>, tensor<i64>, tensor<i64>) -> tensor<4x4xi32>
Traits: AlwaysSpeculatableImplTrait
, InferTensorType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
update | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
start_indices | variadic of 0D tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.einsum
(mhlo::EinsumOp)
Einsum operation
This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/3
Informally, this operation does the same thing as TF's einsum: https://www.tensorflow.org/api_docs/python/tf/einsum
Contoh:
%result = "mhlo.einsum"(%lhs, %rhs) {
einsum_config = "ab,bc->ac"
} : (tensor<4x16xf32>, tensor<16x4xf32>) -> tensor<4x4xf32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
einsum_config | ::mlir::StringAttr | atribut string |
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.erf
(mhlo::ErfOp)
Erf operation
Sintaksis:
operation ::= `mhlo.erf` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise erf operation on operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#erf
Contoh:
%result = mhlo.erf %operand : tensor<2x2xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
mhlo.exponential
(mhlo::ExpOp)
Exp operation
Sintaksis:
operation ::= `mhlo.exponential` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise exponential operation on operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#exponential
Contoh:
%result = mhlo.exponential %operand : tensor<2x2xf64>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
mhlo.exponential_minus_one
(mhlo::Expm1Op)
Expm1 operation
Sintaksis:
operation ::= `mhlo.exponential_minus_one` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise exponential minus one operation on operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#exponential_minus_one
Contoh:
%result = mhlo.exponential_minus_one %operand : tensor<2xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
mhlo.fft
(mhlo::FftOp)
Fft operation
Performs the forward and inverse Fourier transforms for real and complex inputs/outputs.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#fft
Contoh:
%result = mhlo.fft %operand, type = FFT, length = [4] : (tensor<4xcomplex<f32>>) -> tensor<4xcomplex<f32>>
Traits: AlwaysSpeculatableImplTrait
, InferTensorType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
fft_type | ::mlir::mhlo::FftTypeAttr | XLA fast fourier transform type. |
fft_length | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.floor
(mhlo::FloorOp)
Operasi lantai
Sintaksis:
operation ::= `mhlo.floor` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise floor of operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#floor
Contoh:
%result = mhlo.floor %operand : tensor<2xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
mhlo.fusion
(mhlo::FusionOp)
Fusion operation
This operation is private to the XLA compiler, so it is does not yet have a specification.
Informally, this operation consists of a group of basic ops (represented as a region attached to it). It serves as a hint to the backend that it is beneficial to emit the contained ops into a single loop nest or kernel.
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
fusion_kind | ::mlir::mhlo::FusionKindAttr | fusion kind |
output_operand_aliases | ::mlir::ArrayAttr | Aliasing attribute for outputs and operands of Fusion |
Operands:
Operan | Keterangan |
---|---|
inputs | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token |
Hasil:
Hasil | Keterangan |
---|---|
results | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or nested tuple with any combination of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token values |
mhlo.gather
(mhlo::GatherOp)
Gather operation
Gathers slices from operand
tensor from offsets specified in start_indices
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#gather
Contoh:
%result = "mhlo.gather"(%operand, %start_indices) {
dimension_numbers = #mhlo.gather<
offset_dims = [2, 3],
collapsed_slice_dims = [0],
start_index_map = [0, 2],
index_vector_dim = 2>,
slice_sizes = dense<[0, 2, 2]> : tensor<3xi64>,
indices_are_sorted = false
} : (tensor<3x4x2xi32>, tensor<2x3x2xi64>) -> tensor<2x3x2x2xi32>
Traits: AlwaysSpeculatableImplTrait
, InferTensorType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
dimension_numbers | ::mlir::mhlo::GatherDimensionNumbersAttr | Attribute that models the dimension information for gather |
slice_sizes | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
indices_are_sorted | ::mlir::BoolAttr | atribut bool |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
start_indices | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.get_dimension_size
(mhlo::GetDimensionSizeOp)
GetDimensionSize operation
Produces the size of the given dimension
of the operand
.
See https://github.com/openxla/stablehlo/blob/main/docs/spec.md#get_dimension_size
Contoh:
%result = mhlo.get_dimension_size %operand, dim = 1 : (tensor<2x3xf32>) -> tensor<i32>
Traits: AlwaysSpeculatableImplTrait
, InferTensorType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
dimension | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | tensor of 32-bit signless integer values |
mhlo.get_tuple_element
(mhlo::GetTupleElementOp)
GetTupleElement operation
Sintaksis:
operation ::= `mhlo.get_tuple_element` $operand `[` $index `]` attr-dict `:` functional-type(operands, results)
Extracts element at index
position of the operand
tuple and produces a result
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#get_tuple_element
Contoh:
%result = mhlo.get_tuple_element %operand[0] : (tuple<tensor<2xf32>, tuple<tensor<i32>>>) -> tensor<2xf32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
index | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 32-bit |
Operands:
Operan | Keterangan |
---|---|
operand | nested tuple with any combination of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token or nested tuple with any combination of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token values |
mhlo.if
(mhlo::IfOp)
Jika operasi
Produces the output from executing exactly one branch from true_branch
or false_branch
depending on the value of pred
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#if
Example: %result = "mhlo.if"(%pred) ({ "mhlo.return"(%result_true_branch) : (tensor
Traits: RecursiveMemoryEffects
, SingleBlockImplicitTerminator<ReturnOp>
, SingleBlock
Antarmuka: InferTypeOpInterface
Operands:
Operan | Keterangan |
---|---|
pred | ranked tensor of pred (AKA boolean or 1-bit integer) values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token |
mhlo.imag
(mhlo::ImagOp)
Imag operation
Sintaksis:
operation ::= `mhlo.imag` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Extracts the imaginary part, element-wise, from the operand
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#imag
Contoh:
%result = mhlo.imag %operand : (tensor<2xcomplex<f32>>) -> tensor<2xf32>
Traits: AlwaysSpeculatableImplTrait
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
mhlo.infeed
(mhlo::InfeedOp)
Infeed operation
Reads data from the infeed and produces results
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#infeed
Contoh:
%results:2 = "mhlo.infeed"(%token) {
infeed_config = ""
} : (!mhlo.token) -> (tensor<3x3x3xi32>, !mhlo.token)
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
infeed_config | ::mlir::StringAttr | atribut string |
layout | ::mlir::ArrayAttr | atribut susunan |
Operands:
Operan | Keterangan |
---|---|
token | token |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | variadic of statically shaped tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or statically shaped tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token |
mhlo.iota
(mhlo::IotaOp)
Iota operation
Fills an output
tensor with values in increasing order starting from zero along the iota_dimension
dimension.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#iota
Contoh:
%output = mhlo.iota dim = 0 : tensor<4x5xi32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
iota_dimension | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
Hasil:
Hasil | Keterangan |
---|---|
output | statically shaped tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements values |
mhlo.is_finite
(mhlo::IsFiniteOp)
IsFinite operation
Sintaksis:
operation ::= `mhlo.is_finite` $x attr-dict `:` functional-type(operands, results)
Performs element-wise check whether the value in x
is finite (ie is neither +Inf, -Inf, nor NaN) and produces a y
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#is_finite
Contoh:
%y = mhlo.is_finite %x : (tensor<7xf32>) -> tensor<7xi1>
Traits: AlwaysSpeculatableImplTrait
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
x | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
Hasil:
Hasil | Keterangan |
---|---|
y | ranked tensor of pred (AKA boolean or 1-bit integer) values |
mhlo.log
(mhlo::LogOp)
Operasi pencatatan
Sintaksis:
operation ::= `mhlo.log` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise logarithm operation on operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#log
Contoh:
%result = mhlo.log %operand : tensor<2x2xf64>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
mhlo.log_plus_one
(mhlo::Log1pOp)
Log1p operation
Sintaksis:
operation ::= `mhlo.log_plus_one` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise logarithm plus one operation on operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#log_plus_one
Contoh:
%result = mhlo.log_plus_one %operand : tensor<6xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
mhlo.logistic
(mhlo::LogisticOp)
Logistic operation
Sintaksis:
operation ::= `mhlo.logistic` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise logistic operation on operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#logistic
Contoh:
%result = mhlo.logistic %operand : tensor<2x2xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
mhlo.map
(mhlo::MapOp)
Map operation
Applies a map function computation
to inputs
along the dimensions
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#map
Contoh:
%result = "mhlo.map"(%input0, %input1) ({
^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
%0 = mhlo.multiply %arg0, %arg1 : tensor<i32>
mhlo.return %0 : tensor<i32>
}) {
dimensions = dense<[0, 1]> : tensor<2xi64>
} : (tensor<2x2xi32>, tensor<2x2xi32>) -> tensor<2x2xi32>
Traits: InferTensorType
, RecursiveMemoryEffects
, SameOperandsAndResultShape
, SingleBlockImplicitTerminator<ReturnOp>
, SingleBlock
Interfaces: InferShapedTypeOpInterface
, InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
dimensions | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
Operands:
Operan | Keterangan |
---|---|
inputs | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.maximum
(mhlo::MaxOp)
Max operation
Sintaksis:
operation ::= `mhlo.maximum` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise max operation on tensors lhs
and rhs
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#maximum
Contoh:
%result = mhlo.maximum %lhs, %rhs : tensor<4xf32>
Traits: AlwaysSpeculatableImplTrait
, Commutative
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.minimum
(mhlo::MinOp)
Min operation
Sintaksis:
operation ::= `mhlo.minimum` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise min operation on tensors lhs
and rhs
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#minimum
Contoh:
%result = mhlo.minimum %lhs, %rhs : tensor<4xf32>
Traits: AlwaysSpeculatableImplTrait
, Commutative
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.multiply
(mhlo::MulOp)
Mul operation
Sintaksis:
operation ::= `mhlo.multiply` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise product of two tensors lhs
and rhs
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#multiply
Contoh:
%result = mhlo.multiply %lhs, %rhs : tensor<2xi32>
Traits: AlwaysSpeculatableImplTrait
, Commutative
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.negate
(mhlo::NegOp)
Neg operation
Sintaksis:
operation ::= `mhlo.negate` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise negation of operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#negate
Contoh:
%result = mhlo.negate %operand : tensor<2x3xi32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.not
(mhlo::NotOp)
Not operation
Sintaksis:
operation ::= `mhlo.not` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise NOT of tensor operand
of type integer and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#not
Contoh:
%result = mhlo.not %operand : tensor<5x3x1xi1>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
mhlo.optimization_barrier
(mhlo::OptimizationBarrierOp)
OptimizationBarrier operation
Sintaksis:
operation ::= `mhlo.optimization_barrier` attr-dict ($operand^ `:` custom<PairwiseOpType>(type($operand), type($result))):(`(` `)`)?
Ensures that the operations that produce the operand
are executed before any operations that depend on the result
and prevents compiler transformations from moving operations across the barrier. Other than that, the operation is an identity, ie result
= operand
.
See https://github.com/openxla/stablehlo/blob/main/docs/spec.md#optimization_barrier
Contoh:
%result0, %result1 = mhlo.optimization_barrier %operand0, %operand1 : tensor<f32>, tensor<f32>
Traits: AlwaysSpeculatableImplTrait
, HLO_PairwiseSameOperandAndResultType
Antarmuka: ConditionallySpeculatable
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token |
Hasil:
Hasil | Keterangan |
---|---|
result | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token |
mhlo.or
(mhlo::OrOp)
Or operation
Sintaksis:
operation ::= `mhlo.or` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise OR of two tensors lhs
and rhs
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#or
Contoh:
%result = mhlo.or %lhs, %rhs : tensor<2xi1>
Traits: AlwaysSpeculatableImplTrait
, Commutative
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
rhs | ranked tensor of pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.outfeed
(mhlo::OutfeedOp)
Outfeed operation
Writes inputs
to the outfeed and produces a result
token.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#outfeed
Contoh:
%result = "mhlo.outfeed"(%input0, %token) {
outfeed_config = ""
} : (tensor<3x3x3xi32>, !mhlo.token) -> !mhlo.token
Antarmuka: InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
outfeed_config | ::mlir::StringAttr | atribut string |
Operands:
Operan | Keterangan |
---|---|
inputs | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
token | token |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | token |
mhlo.pad
(mhlo::PadOp)
Pad operation
Expands operand
by padding around the tensor as well as between the elements of the tensor with the given padding_value
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#pad
Contoh:
%0 = mhlo.pad %arg0, %arg1, low = [0, 1], high = [2, 1], interior = [1, 2]
: (tensor<2x3xi32>, tensor<i32>) -> tensor<5x9xi32>
Ciri-ciri: AlwaysSpeculatableImplTrait
, SameOperandsAndResultElementType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
edge_padding_low | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
edge_padding_high | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
interior_padding | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
padding_value | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.partition_id
(mhlo::PartitionIdOp)
PartitionId operation
Sintaksis:
operation ::= `mhlo.partition_id` attr-dict `:` type(results)
Produces partition_id
of the current process.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#partition_id
Contoh:
%result = mhlo.partition_id : tensor<ui32>
Antarmuka: InferTypeOpInterface
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of 32-bit unsigned integer values |
mhlo.popcnt
(mhlo::PopulationCountOp)
PopulationCount operation
Sintaksis:
operation ::= `mhlo.popcnt` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise count of the number of bits set in the operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#popcnt
Contoh:
%result = mhlo.popcnt %operand : tensor<4xi8>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
mhlo.power
(mhlo::PowOp)
Pow operation
Sintaksis:
operation ::= `mhlo.power` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise exponentiation of lhs
tensor by rhs
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#power
Contoh:
%result = mhlo.power %lhs, %rhs : tensor<6xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.real
(mhlo::RealOp)
Real operation
Sintaksis:
operation ::= `mhlo.real` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Extracts the real part, element-wise, from the operand
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#real
Contoh:
%result = mhlo.real %operand : (tensor<2xcomplex<f32>>) -> tensor<2xf32>
Traits: AlwaysSpeculatableImplTrait
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
mhlo.real_dynamic_slice
(mhlo::RealDynamicSliceOp)
RealDynamicSlice operation
Sintaksis:
operation ::= `mhlo.real_dynamic_slice` operands attr-dict `:` functional-type(operands, results)
This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8
Informally, this operation does the same thing as SliceOp except that start_indices
, limit_indices
and strides
are specified dynamically: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#slice
Contoh:
%result = mhlo.real_dynamic_slice %operand,
%start_indices, %limit_indices, %strides
: (tensor<256x?xf32>, tensor<2xindex>, tensor<2xindex>, tensor<2xindex>) -> tensor<256x?xf32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
start_indices | 1D tensor of index or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
limit_indices | 1D tensor of index or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
strides | 1D tensor of index or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.recv
(mhlo::RecvOp)
Recv operation
Receives data from a channel with channel_id
and produces results
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#recv
Contoh:
%results:2 = "mhlo.recv"(%token) {
// channel_id = 5 : i64,
// channel_type = #stablehlo<channel_type HOST_TO_DEVICE>,
channel_handle = #mhlo.channel_handle<handle = 5, type = 3>,
is_host_transfer = true
} : (!mhlo.token) -> (tensor<3x4xi32>, !mhlo.token)
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
channel_handle | ::mlir::mhlo::ChannelHandleAttr | two 64-bit integers 'handle' and 'type' |
is_host_transfer | ::mlir::BoolAttr | atribut bool |
Operands:
Operan | Keterangan |
---|---|
token | token |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | variadic of statically shaped tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or statically shaped tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token |
mhlo.reduce
(mhlo::ReduceOp)
Reduce operation
Applies a reduction function body
to inputs
and init_values
along the dimensions
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce
Contoh:
%result = "mhlo.reduce"(%input, %init_value) ({
^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
%0 = "mhlo.add"(%arg0, %arg1) : (tensor<i32>, tensor<i32>) -> tensor<i32>
"mhlo.return"(%0) : (tensor<i32>) -> ()
}) {
dimensions = dense<1> : tensor<1xi64>
} : (tensor<1x6xi32>, tensor<i32>) -> tensor<1xi32>
Traits: InferTensorType
, RecursiveMemoryEffects
, SameVariadicOperandSize
, SingleBlockImplicitTerminator<ReturnOp>
, SingleBlock
Interfaces: InferShapedTypeOpInterface
, InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
dimensions | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
Operands:
Operan | Keterangan |
---|---|
inputs | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
init_values | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
mhlo.reduce_precision
(mhlo::ReducePrecisionOp)
ReducePrecision operation
Sintaksis:
operation ::= `mhlo.reduce_precision` $operand `,` `format` `=` custom<ExponentMantissa>($exponent_bits, $mantissa_bits)
attr-dict `:` custom<SameOperandsAndResultType>(type($operand), type($output))
Performs element-wise conversion of operand
to another floating-point type that uses exponent_bits
and mantissa_bits
and back to the original floating-point type and produces an output
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce_precision
Contoh:
%output = mhlo.reduce_precision %operand, format = e5m2 : tensor<6xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
exponent_bits | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 32-bit |
mantissa_bits | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 32-bit |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
Hasil:
Hasil | Keterangan |
---|---|
output | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
mhlo.reduce_scatter
(mhlo::ReduceScatterOp)
Operasi ReduceScatter
Within each process group in the process grid, performs reduction, using computations
, over the values of the operand
tensor from each process, splits the reduction result along scatter_dimension
into parts, and scatters the split parts between the processes to produce the result
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce_scatter
Contoh:
%result = "mhlo.reduce_scatter"(%operand) ({
^bb0(%arg0: tensor<f32>, %arg1: tensor<f32>):
%0 = mhlo.add %arg0, %arg1 : tensor<f32>
mhlo.return %0 : tensor<f32>
}) {
scatter_dimension = 1 : i64,
replica_groups = dense<[[0, 1]]> : tensor<1x2xi64>,
// channel_id = 0
channel_handle = #mhlo.channel_handle<handle = 0, type = 0>
// use_global_device_ids = false
} : (tensor<2x4xf32>) -> tensor<2x2xf32>
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
scatter_dimension | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
replica_groups | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
channel_handle | ::mlir::mhlo::ChannelHandleAttr | two 64-bit integers 'handle' and 'type' |
use_global_device_ids | ::mlir::UnitAttr | atribut satuan |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.reduce_window
(mhlo::ReduceWindowOp)
ReduceWindow operation
Applies a reduction function body
to windows of inputs
and init_values
and produces results
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reduce_window
Contoh:
%result = "mhlo.reduce_window"(%input, %init_value) ({
^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
%0 = mhlo.add %arg0, %arg1 : tensor<i32>
mhlo.return %0 : tensor<i32>
}) {
window_dimensions = dense<[2, 1]> : tensor<2xi64>,
window_strides = dense<[4, 1]> : tensor<2xi64>,
base_dilations = dense<[2, 1]> : tensor<2xi64>,
window_dilations = dense<[3, 1]> : tensor<2xi64>,
padding = dense<[[2, 1], [0, 0]]> : tensor<2x2xi64>
} : (tensor<3x2xi32>, tensor<i32>) -> tensor<2x2xi32>
Traits: InferTensorType
, RecursiveMemoryEffects
, SameVariadicOperandSize
, SingleBlockImplicitTerminator<ReturnOp>
, SingleBlock
Interfaces: InferShapedTypeOpInterface
, InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
window_dimensions | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
window_strides | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
base_dilations | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
window_dilations | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
padding | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
Operands:
Operan | Keterangan |
---|---|
inputs | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
init_values | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
mhlo.remainder
(mhlo::RemOp)
Rem operation
Sintaksis:
operation ::= `mhlo.remainder` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise remainder of dividend lhs
and divisor rhs
tensors and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#remainder
Contoh:
%result = mhlo.remainder %lhs, %rhs : tensor<4xi64>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.replica_id
(mhlo::ReplicaIdOp)
ReplicaId operation
Sintaksis:
operation ::= `mhlo.replica_id` attr-dict `:` type(results)
Produces replica_id
of the current process.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#replica_id
Contoh:
%result = mhlo.replica_id : tensor<ui32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of 32-bit unsigned integer values |
mhlo.reshape
(mhlo::ReshapeOp)
Reshape operation
Sintaksis:
operation ::= `mhlo.reshape` operands attr-dict `:` functional-type(operands, results)
Performs reshape of operand
tensor to a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reshape
Contoh:
%result = mhlo.reshape %operand : (tensor<2xf32>) -> tensor<1x2xf32>
Traits: AlwaysSpeculatableImplTrait
, HLO_CompatibleOperandsAndResultElementType
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | statically shaped tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.return
(mhlo::ReturnOp)
_This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/425
Informally, this operation serves as a terminator for regions defined by
the StableHLO ops. Non-StableHLO ops, e.g. `func.func`, have their own
terminators, e.g. `func.return`.
Example:
```mlir
%result = "mhlo.reduce"(%input, %init_value) ({
^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
%0 = "mhlo.add"(%arg0, %arg1) : (tensor<i32>, tensor<i32>) -> tensor<i32>
"mhlo.return"(%0) : (tensor<i32>) -> ()
}) {
dimensions = dense<1> : tensor<1xi64>
} : (tensor<1x6xi32>, tensor<i32>) -> tensor<1xi32>
```_
Syntax:
```
operation ::= mhlo.return
$results attr-dict ( :
type($results)^)?
Traits: `AlwaysSpeculatableImplTrait`, `Terminator`
Interfaces: `ConditionallySpeculatable`, `NoMemoryEffect (MemoryEffectOpInterface)`
Effects: `MemoryEffects::Effect{}`
#### Operands:
| Operand | Description |
| :-----: | ----------- |
| `results` | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token or nested tuple with any combination of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token values
### `mhlo.reverse` (mhlo::ReverseOp)
_Reverse operation_
Reverses the order of elements in the `operand` along the specified
`dimensions` and produces a `result` tensor.
See:
<a href="https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reverse">https://github.com/openxla/stablehlo/blob/main/docs/spec.md#reverse</a>
Example:
```mlir
%result = mhlo.reverse %operand, dims = [1] : tensor<3x2xi32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
dimensions | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.rng
(mhlo::RngOp)
Rng operation
Generates random numbers using the rng_distribution
algorithm and produces a result
tensor of a given shape shape
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#rng
Contoh:
%result = mhlo.rng %a, %b, %shape, distribution = NORMAL : (tensor<i32>, tensor<i32>, tensor<2xi64>) -> tensor<3x3xi32>
Traits: InferTensorType
Interfaces: InferShapedTypeOpInterface
, InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
rng_distribution | ::mlir::mhlo::RngDistributionAttr | XLA PRNG distribution to be used. |
Operands:
Operan | Keterangan |
---|---|
a | 0D tensor of pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
b | 0D tensor of pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
shape | 1D tensor of index or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
mhlo.rng_bit_generator
(mhlo::RngBitGeneratorOp)
RngBitGenerator operation
Returns an output
filled with uniform random data and an updated output state output_state
given an initial state initial_state
using the pseudorandom number generator algorithm rng_algorithm
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#rng_bit_generator
Contoh:
%output_state, %output = mhlo.rng_bit_generator %initial_state, algorithm = THREE_FRY : (tensor<2xui64>) -> (tensor<2xui64>, tensor<2x2xui64>)
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
rng_algorithm | ::mlir::mhlo::RngAlgorithmAttr | XLA PRNG algorithm to be used. |
Operands:
Operan | Keterangan |
---|---|
initial_state | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
Hasil:
Hasil | Keterangan |
---|---|
output_state | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
output | statically shaped tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
mhlo.round_nearest_afz
(mhlo::RoundOp)
Round operation
Sintaksis:
operation ::= `mhlo.round_nearest_afz` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise rounding towards the nearest integer, breaking ties away from zero, on the operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#round_nearest_afz
Contoh:
%result = mhlo.round_nearest_afz %operand : tensor<5xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
mhlo.round_nearest_even
(mhlo::RoundNearestEvenOp)
RoundNearestEven operation
Sintaksis:
operation ::= `mhlo.round_nearest_even` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise rounding towards the nearest integer, breaking ties towards the even integer, on the operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#round_nearest_even
Contoh:
%result = mhlo.round_nearest_even %operand : tensor<5xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
mhlo.rsqrt
(mhlo::RsqrtOp)
Rsqrt operation
Sintaksis:
operation ::= `mhlo.rsqrt` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise reciprocal square root operation on operand
tensor and produces a result
tensor, implementing the rSqrt
operation from the IEEE-754 specification.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#rsqrt
Contoh:
%result = mhlo.rsqrt %operand : tensor<2x2xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
mhlo.scatter
(mhlo::ScatterOp)
Scatter operation
Produces results
tensors which are equal to inputs
tensors except that several slices specified by scatter_indices
are updated with the values updates
using update_computation
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#scatter
Contoh:
%result = "mhlo.scatter"(%input, %scatter_indices, %update) ({
^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
%0 = mhlo.add %arg0, %arg1 : tensor<i32>
mhlo.return %0 : tensor<i32>
}) {
scatter_dimension_numbers = #mhlo.scatter<
update_window_dims = [2,3],
inserted_window_dims = [0],
scatter_dims_to_operand_dims = [1, 0],
index_vector_dim = 2>,
indices_are_sorted = false,
unique_indices = false
} : (tensor<3x4x2xi32>, tensor<2x3x2xi64>, tensor<2x3x2x2xi32>) -> tensor<3x4x2xi32>
Traits: RecursiveMemoryEffects
, SameVariadicOperandSize
Antarmuka: InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
scatter_dimension_numbers | ::mlir::mhlo::ScatterDimensionNumbersAttr | Attribute that models the dimension information for scatter |
indices_are_sorted | ::mlir::BoolAttr | atribut bool |
unique_indices | ::mlir::BoolAttr | atribut bool |
Operands:
Operan | Keterangan |
---|---|
inputs | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
scatter_indices | ranked tensor of integer or index values |
updates | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
mhlo.select
(mhlo::SelectOp)
Pilih operasi
Sintaksis:
operation ::= `mhlo.select` operands attr-dict `:`
custom<SelectOpType>(type($pred), type($on_true), type($on_false), type($result))
Produces a result
tensor where each element is selected from on_true
or on_false
tensor based on the value of the corresponding element of pred
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#select
Contoh:
%result = mhlo.select %pred, %on_true, %on_false : tensor<2x2xi1>, tensor<2x2xi32>
Traits: AlwaysSpeculatableImplTrait
, HLO_BroadcastingElementwise
, InferTensorType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
pred | ranked tensor of pred (AKA boolean or 1-bit integer) values |
on_true | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
on_false | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.select_and_scatter
(mhlo::SelectAndScatterOp)
SelectAndScatter operation
Scatters the values from the source
tensor using scatter
based on the outcome of reduce_window
of the input
tensor using select
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#select_and_scatter
Contoh:
%result = "mhlo.select_and_scatter"(%operand, %source, %init_value) ({
^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
%0 = "mhlo.compare"(%arg0, %arg1) {
comparison_direction = #stablehlo<comparison_direction GE>
} : (tensor<i32>, tensor<i32>) -> tensor<i1>
"mhlo.return"(%0) : (tensor<i1>) -> ()
}, {
^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
%0 = "mhlo.add"(%arg0, %arg1) : (tensor<i32>, tensor<i32>) -> tensor<i32>
"mhlo.return"(%0) : (tensor<i32>) -> ()
}) {
window_dimensions = dense<[3, 1]> : tensor<2xi64>,
window_strides = dense<[2, 1]> : tensor<2xi64>,
padding = dense<[[0, 1], [0, 0]]> : tensor<2x2xi64>
} : (tensor<4x2xi32>, tensor<2x2xi32>, tensor<i32>) -> tensor<4x2xi32>
Traits: RecursiveMemoryEffects
Antarmuka: InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
window_dimensions | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
window_strides | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
padding | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
source | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
init_value | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.send
(mhlo::SendOp)
Send operation
Sends inputs
to a channel channel_id
and produces a result
token.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#send
Contoh:
%result = "mhlo.send"(%operand, %token) {
// channel_id = 5 : i64,
// channel_type = #stablehlo<channel_type DEVICE_TO_HOST>,
channel_handle = #mhlo.channel_handle<handle = 5, type = 2>,
is_host_transfer = true
} : (tensor<3x4xi32>, !mhlo.token) -> !mhlo.token
Antarmuka: InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
channel_handle | ::mlir::mhlo::ChannelHandleAttr | two 64-bit integers 'handle' and 'type' |
is_host_transfer | ::mlir::BoolAttr | atribut bool |
Operands:
Operan | Keterangan |
---|---|
inputs | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
token | token |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | token |
mhlo.set_dimension_size
(mhlo::SetDimensionSizeOp)
SetDimensionSize operation
This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/8
Informally, this operation does the same thing as XLA's SetDimensionSize: https://www.tensorflow.org/xla/operation_semantics#setdimensionsize
Contoh:
%0 = mhlo.set_dimension_size %arg0, %arg1, dim = 1 : (tensor<4x2xf32>, tensor<i32>) -> tensor<4x2xf32>
Traits: AlwaysSpeculatableImplTrait
, InferTensorType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
dimension | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
size | tensor of 32-bit signless integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.shift_left
(mhlo::ShiftLeftOp)
ShiftLeft operation
Sintaksis:
operation ::= `mhlo.shift_left` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise left-shift operation on the lhs
tensor by rhs
number of bits and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#shift_left
Contoh:
%result = mhlo.shift_left %lhs, %rhs : tensor<6xi8>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
rhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
mhlo.shift_right_arithmetic
(mhlo::ShiftRightArithmeticOp)
ShiftRightArithmetic operation
Sintaksis:
operation ::= `mhlo.shift_right_arithmetic` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise arithmetic right-shift operation on the lhs
tensor by rhs
number of bits and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#shift_right_arithmetic
Contoh:
%result = mhlo.shift_right_arithmetic %lhs, %rhs : tensor<6xi8>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
rhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
mhlo.shift_right_logical
(mhlo::ShiftRightLogicalOp)
ShiftRightLogical operation
Sintaksis:
operation ::= `mhlo.shift_right_logical` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise logical right-shift operation on the lhs
tensor by rhs
number of bits and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#shift_right_logical
Contoh:
%result = mhlo.shift_right_logical %lhs, %rhs : tensor<6xi8>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
rhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
mhlo.sign
(mhlo::SignOp)
Sign operation
Sintaksis:
operation ::= `mhlo.sign` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Returns the sign of the operand
element-wise and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sign
Contoh:
%result = mhlo.sign %operand : tensor<7xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of 4/8/16/32/64-bit signless integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of 4/8/16/32/64-bit signless integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
mhlo.sine
(mhlo::SineOp)
Sine operation
Sintaksis:
operation ::= `mhlo.sine` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise sine operation on operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sine
Contoh:
%result = mhlo.sine %operand : tensor<2xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
mhlo.slice
(mhlo::SliceOp)
Operasi irisan
Extracts a slice from the operand
using statically-computed starting indices and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#slice
Contoh:
%result = "mhlo.slice" (%operand) {
start_indices = dense<[1, 2]> : tensor<2xi64>,
limit_indices = dense<[3, 4]> : tensor<2xi64>,
strides = dense<1> : tensor<2xi64>
} : (tensor<3x4xi64>) -> tensor<2x2xi64>
Ciri-ciri: AlwaysSpeculatableImplTrait
, SameOperandsAndResultElementType
Antarmuka: ConditionallySpeculatable
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
start_indices | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
limit_indices | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
strides | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.sort
(mhlo::SortOp)
Sort operation
Sorts a variadic number of tensors in inputs
together, according to a custom comparator
, along the given dimension
and produces a variadic number of tensors as results
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sort
Contoh:
%result0, %result1 = "mhlo.sort"(%input0, %input1) ({
^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>, %arg2: tensor<i32>, %arg3: tensor<i32>):
%predicate = "mhlo.compare"(%arg0, %arg1) {
comparison_direction = #stablehlo<comparison_direction GT>
} : (tensor<i32>, tensor<i32>) -> tensor<i1>
"mhlo.return"(%predicate) : (tensor<i1>) -> ()
}) {
dimension = 0 : i64,
is_stable = true
} : (tensor<2x3xi32>, tensor<2x3xi32>) -> (tensor<2x3xi32>, tensor<2x3xi32>)
Traits: InferTensorType
, RecursiveMemoryEffects
, SameOperandsAndResultShape
Interfaces: InferShapedTypeOpInterface
, InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
dimension | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
is_stable | ::mlir::BoolAttr | atribut bool |
Operands:
Operan | Keterangan |
---|---|
inputs | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
mhlo.sparse_dot
(mhlo::SparseDotOp)
Sparse dot operation
Similar to dot_general
operation, with one or both of the operands being sparse. An additional argument provides sparsity meta information. Disclaimer: this op is experimental / a work in progress.
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
lhs_sparsity | ::mlir::mhlo::SparsityDescriptorAttr | Describes structured (N:M) sparsity configuration |
rhs_sparsity | ::mlir::mhlo::SparsityDescriptorAttr | Describes structured (N:M) sparsity configuration |
dot_dimension_numbers | ::mlir::mhlo::DotDimensionNumbersAttr | Attribute that models the dimension information for dot. |
precision_config | ::mlir::ArrayAttr | Precision Config attribute |
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
meta | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4 /8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer nilai-nilai |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.sqrt
(mhlo::SqrtOp)
Sqrt operation
Sintaksis:
operation ::= `mhlo.sqrt` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise square root operation on operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#sqrt
Contoh:
%result = mhlo.sqrt %operand : tensor<2x2xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
mhlo.stochastic_convert
(mhlo::StochasticConvertOp)
StochasticConvert operation
This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/295
Informally, this operation performs element-wise conversion of values from a bigger type to a smaller one with stochastic rounding using the random number passed in.
Traits: AlwaysSpeculatableImplTrait
, Elementwise
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
random | ranked tensor of 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.subtract
(mhlo::SubtractOp)
Subtract operation
Sintaksis:
operation ::= `mhlo.subtract` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise subtraction of two tensors lhs
and rhs
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#subtract
Contoh:
%result = mhlo.subtract %lhs, %rhs : tensor<2xi32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
rhs | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.tan
(mhlo::TanOp)
Tan operation
Sintaksis:
operation ::= `mhlo.tan` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
This operation is a work in progress, so it is not yet included in the specification: https://github.com/openxla/stablehlo/issues/954
Informally, this operation returns Tan(operand)
element-wise.
Contoh:
%0 = mhlo.tan %arg0 : tensor<2xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements values |
mhlo.tanh
(mhlo::TanhOp)
Tanh operation
Sintaksis:
operation ::= `mhlo.tanh` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise hyperbolic tangent operation on operand
tensor and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#tanh
Contoh:
%result = mhlo.tanh %operand : tensor<2xf32>
Traits: AlwaysSpeculatableImplTrait
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values |
mhlo.topk
(mhlo::TopKOp)
TopK operation
Sintaksis:
operation ::= `mhlo.topk` `(`$operand `,` `k` `=` $k (`,` `largest` `=` $largest^)? `)` attr-dict `:`
type($operand) `->` `(`type($values)`,` type($indices)`)`
Returns top k
values and their indices, along the last dimension of the operand if largest=true
or the bottom k
values if largest=false
.
See: https://www.tensorflow.org/xla/operation_semantics#top-k
Contoh:
%values, %indices = mhlo.topk(%operand, k=5, largest=true)
: tensor<100xf32> -> (tensor<5xf32>, tensor<5xi32>)
Traits: InferTensorType
, RecursiveMemoryEffects
Interfaces: InferShapedTypeOpInterface
, InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
k | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
largest | ::mlir::BoolAttr | atribut bool |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
values | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
indices | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.torch_index_select
(mhlo::TorchIndexSelectOp)
TorchIndexSelect operation
This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/3
Informally, this operation does the same thing as PyTorch's index_select, augmented with support for batch dimensions: https://pytorch.org/docs/stable/generated/torch.index_select.html
The batch_dims
attribute specifies the number of major batch dimensions (0 or more) that act like a multidimensional loop over both the operand and the index.
Contoh:
%result = "mhlo.torch_index_select"(%operand, %index) {
dim = 2 : i64,
batch_dims = 1 : i64
} : (tensor<8x128x3072x64xf32>, tensor<8x16x1024xi32>) -> tensor<8x128x16x1024x64xf32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
dim | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
batch_dims | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
index | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.trace
(mhlo::TraceOp)
Trace operation
Sintaksis:
operation ::= `mhlo.trace` $operand `,` $tag attr-dict `:` type($operand)
This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/604
It is not used by JAX, PyTorch or TensorFlow, so it looks like we should've classified it as "Private to XLA" and not included it in StableHLO in the first place. With that in mind, its semantics will not be documented here.
Contoh:
mhlo.trace %arg0, "In test code." : tensor<5x1x5xi32>
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
tag | ::mlir::StringAttr | atribut string |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.transpose
(mhlo::TransposeOp)
Transpose operation
Permutes the dimensions of operand
tensor using permutation
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#transpose
Contoh:
%0 = mhlo.transpose %arg0, dims = [2, 1, 0] : (tensor<1x2x3xi32>) -> tensor<3x2x1xi32>
Traits: AlwaysSpeculatableImplTrait
, HLO_CompatibleOperandsAndResultElementType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
permutation | ::mlir::DenseIntElementsAttr | 64-bit signless integer elements attribute |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.triangular_solve
(mhlo::TriangularSolveOp)
TriangularSolve operation
Solves batches of systems of linear equations with lower or upper triangular coefficient matrices.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#triangular_solve
Contoh:
%result = "mhlo.triangular_solve"(%a, %b) {
left_side = true,
lower = true,
unit_diagonal = false,
transpose_a = #stablehlo<transpose NO_TRANSPOSE>
} : (tensor<3x3xf32>, tensor<3x3xf32>) -> tensor<3x3xf32>
Traits: AlwaysSpeculatableImplTrait
, InferTensorType
, SameOperandsAndResultElementType
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
left_side | ::mlir::BoolAttr | atribut bool |
lower | ::mlir::BoolAttr | atribut bool |
unit_diagonal | ::mlir::BoolAttr | atribut bool |
transpose_a | ::mlir::mhlo::TransposeAttr | Transpose options |
Operands:
Operan | Keterangan |
---|---|
a | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements values |
b | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or complex type with 32-bit float or 64-bit float elements values |
mhlo.tuple
(mhlo::TupleOp)
Tuple operation
Sintaksis:
operation ::= `mhlo.tuple` $val attr-dict `:` custom<TupleOpType>(type($val), type($result))
Produces a result
tuple from values val
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#tuple
Contoh:
%result = mhlo.tuple %val0, %val1 : tuple<tensor<2xf32>, tuple<tensor<i32>>>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
val | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token or nested tuple with any combination of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token values |
Hasil:
Hasil | Keterangan |
---|---|
result | nested tuple with any combination of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token values |
mhlo.unary_einsum
(mhlo::UnaryEinsumOp)
UnaryEinsum operation
This operation is on its way out of StableHLO, so it is not included in the specification: https://github.com/openxla/stablehlo/issues/3
Informally, this operation does the same thing as TF's einsum: https://www.tensorflow.org/api_docs/python/tf/einsum
Contoh:
%result = "mhlo.unary_einsum"(%operand) {
einsum_config = "ab->a"
} : (tensor<4x16xf32>) -> tensor<4xf32>
Ciri-ciri: AlwaysSpeculatableImplTrait
Antarmuka: ConditionallySpeculatable
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
einsum_config | ::mlir::StringAttr | atribut string |
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.uniform_dequantize
(mhlo::UniformDequantizeOp)
UniformDequantize operation
Sintaksis:
operation ::= `mhlo.uniform_dequantize` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise conversion of quantized tensor operand
to a floating-point tensor result
according to the quantization parameters defined by the operand
type.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#uniform_dequantize
Contoh:
%result = mhlo.uniform_dequantize %operand : (tensor<16x16x!quant.uniform<i8:f32, 34.0:16>>) -> tensor<16x16xf32>
Traits: AlwaysSpeculatableImplTrait
, Elementwise
, InferTensorType
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type values |
mhlo.uniform_quantize
(mhlo::UniformQuantizeOp)
UniformQuantize operation
Sintaksis:
operation ::= `mhlo.uniform_quantize` $operand attr-dict
`:` custom<SameOperandsAndResultType>(type($operand), type($result))
Performs element-wise conversion of floating-point tensor or quantized tensor operand
to a quantized tensor result
according to the quantization parameters defined by the result
type.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#uniform_quantize
Contoh:
%result = mhlo.uniform_quantize %operand : (tensor<16x16xf32>) -> tensor<16x16x!quant.uniform<ui8:f32, 34.0:16>>
Traits: AlwaysSpeculatableImplTrait
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
operand | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
mhlo.while
(mhlo::WhileOp)
While operation
Produces the output from executing body
function 0 or more times while the cond
function outputs true
.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#while
Contoh:
%results0, %results1 = "mhlo.while"(%operand0, %operand1) ({
^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
%0 = "mhlo.compare"(%arg0, %arg1) {
comparison_direction = #stablehlo<comparison_direction LT>
} : (tensor<i32>, tensor<i32>) -> tensor<i1>
"mhlo.return"(%0) : (tensor<i1>) -> ()
}, {
^bb0(%arg0: tensor<i32>, %arg1: tensor<i32>):
%0 = "mhlo.add"(%arg0, %constant0) : (tensor<i32>, tensor<i32>) -> tensor<i32>
"mhlo.return"(%0, %arg1) : (tensor<i32>, tensor<i32>) -> ()
}) : (tensor<i32>, tensor<i32>) -> (tensor<i32>, tensor<i32>)
Traits: RecursiveMemoryEffects
, SingleBlockImplicitTerminator<ReturnOp>
, SingleBlock
Interfaces: InferTypeOpInterface
, OpAsmOpInterface
Operands:
Operan | Keterangan |
---|---|
operand | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | variadic of ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer values or ranked tensor of 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values or token |
mhlo.xla.rng_get_and_update_state
(mhlo::XlaRngGetAndUpdateStateOp)
XlaRngGetAndUpdateState operation
Sintaksis:
operation ::= `mhlo.xla.rng_get_and_update_state` attr-dict
This operation is private to the XLA compiler, so it is does not yet have a specification.
Informally, this operation represents the change of the global random number generator state for rng instructions. The global state is incremented by delta and the old state is returned.
The output is currently defined for a single output type. If this changes in the future to support multiple types, lowering to use of a global memref must ensure that a single memref is still used and updated appropriately.
Antarmuka: InferTypeOpInterface
Atribut:
Atribut | Tipe MLIR | Keterangan |
---|---|---|
delta | ::mlir::IntegerAttr | Atribut bilangan bulat tanpa tanda 64-bit |
Hasil:
Hasil | Keterangan |
---|---|
«unnamed» | statically shaped tensor of 64-bit unsigned integer values |
mhlo.xor
(mhlo::XorOp)
Xor operation
Sintaksis:
operation ::= `mhlo.xor` $lhs `,` $rhs attr-dict
`:` custom<SameOperandsAndResultType>(type($lhs), type($rhs), type($result))
Performs element-wise XOR of two tensors lhs
and rhs
and produces a result
tensor.
See: https://github.com/openxla/stablehlo/blob/main/docs/spec.md#xor
Contoh:
%result = mhlo.xor %lhs, %rhs : tensor<2xi32>
Traits: AlwaysSpeculatableImplTrait
, Commutative
, CompatibleOperandsAndResultType
, Elementwise
, SameOperandsAndResultShape
Interfaces: ConditionallySpeculatable
, InferShapedTypeOpInterface
, InferTypeOpInterface
, NoMemoryEffect (MemoryEffectOpInterface)
Efek: MemoryEffects::Effect{}
Operands:
Operan | Keterangan |
---|---|
lhs | ranked tensor of pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
rhs | ranked tensor of pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer values |
Hasil:
Hasil | Keterangan |
---|---|
result | ranked tensor of f8E4M3B11FNUZ type or f8E4M3FN type or f8E4M3FNUZ type or f8E5M2 type or f8E5M2FNUZ type or 16-bit float or 32-bit float or 64-bit float or bfloat16 type or pred (AKA boolean or 1-bit integer) or 4/8/16/32/64-bit signless integer or 4/8/16/32/64-bit unsigned integer or complex type with 32-bit float or 64-bit float elements or 4/8/16/32-bit uniform quantized signed integer or 4/8/16/32-bit uniform quantized unsigned integer or 4/8/16/32-bit uniform quantized per axis signed integer or 4/8/16/32-bit uniform quantized per axis unsigned integer values |
Atribut
ArgResultAliasAttr
Attribute that models the alias relationship of entry function argument
This attribute captures the alias relationship of an MHLO main function argument to one of the results, denoted by resultIndex
. The argTupleIndices
and resultTupleIndices
are used to index into nested tuples in operand and result respectively. If isMustAlias
is true then the operand-result pair must alias.
This is meant to be used as an attribute on a function argument in MHLO. For example, in the following code it expresses that %arg1
may alias 0-th result.
func @main(%arg0: tensor<2xf32>, %arg1: tensor<3xf32> {mhlo.result_alias =
mhlo.result_alias<result_index = [2], ...>}
) -> tensor<2xf32>, tensor<3xf32> {
// function body ...
}
Parameter:
Parameter | tipe C++ | Keterangan |
---|---|---|
argTupleIndices | ::llvm::ArrayRef<int64_t> | Dimensi |
indeks hasil | int64_t | |
resultTupleIndices | ::llvm::ArrayRef<int64_t> | Dimensi |
isMustAlias | bool |
ChannelHandleAttr
two 64-bit integers 'handle' and 'type'
Sintaksis:
#mhlo.channel_handle<
int64_t, # handle
int64_t # type
>
Parameter:
Parameter | tipe C++ | Keterangan |
---|---|---|
menangani | int64_t | |
jenis | int64_t |
ComparisonDirectionAttr
Which comparison operation to perform.
Sintaksis:
#mhlo.comparison_direction<
::mlir::mhlo::ComparisonDirection # value
>
Kasus enum:
- EQ (
EQ
) - NE (
NE
) - GE (
GE
) - GT (
GT
) - LE (
LE
) - LT (
LT
) #### Parameters:
Parameter | tipe C++ | Keterangan |
---|---|---|
nilai | ::mlir::mhlo::ComparisonDirection | an enum of type ComparisonDirection |
ComparisonTypeAttr
Which comparison type to use.
Sintaksis:
#mhlo.comparison_type<
::mlir::mhlo::ComparisonType # value
>
Kasus enum:
- NOTYPE (
NOTYPE
) - FLOAT (
FLOAT
) - TOTALORDER (
TOTALORDER
) - SIGNED (
SIGNED
) - UNSIGNED (
UNSIGNED
) #### Parameters:
Parameter | tipe C++ | Keterangan |
---|---|---|
nilai | ::mlir::mhlo::ComparisonType | an enum of type ComparisonType |
ConvDimensionNumbersAttr
Structure of dimension information for conv op
Parameter:
Parameter | tipe C++ | Keterangan |
---|---|---|
inputBatchDimension | int64_t | |
inputFeatureDimension | int64_t | |
inputSpatialDimensions | ::llvm::ArrayRef<int64_t> | Dimensi |
kernelInputFeatureDimension | int64_t | |
kernelOutputFeatureDimension | int64_t | |
kernelSpatialDimensions | ::llvm::ArrayRef<int64_t> | Dimensi |
outputBatchDimension | int64_t | |
outputFeatureDimension | int64_t | |
outputSpatialDimensions | ::llvm::ArrayRef<int64_t> | Dimensi |
CrossProgramPrefetchAttr
Argument that is prefetched from another program
Sintaksis:
#mhlo.cross_program_prefetch<
int64_t, # parameter
::llvm::ArrayRef<int64_t>, # indices
std::optional<int64_t> # offset
>
This attribute captures an argument that is prefetched from another program. For a given CrossProgramPrefetchAttr
, parameter
tells us which argument of the main
function of the module is prefetched, and indices
is a shape index telling us what subshape of that argument is prefetched.
A shape has a subshape iff it is a tuple. In that case, the subshape of the tuple by indices
is the shape achieved after indexing by each element of indices
in turn. For example, the [1,0] subshape of tuple<tuple<token, token>, tuple<tensor<i32>, token>>
is tensor<i32>
.
An empty value for indices
means the whole shape is prefetched.
Misalnya,
module attributes { mhlo.cross_program_prefetch = [ #mhlo.cross_program_prefetch< parameter = 0, indices = [0]> ]} {
func.func @copy(%arg0 : tuple<tensor<2x3xi32>, tensor<i32>>) -> tuple<tensor<2x3xi32>, tensor<i32>> {
%0 = "mhlo.copy"(%arg0) {is_cross_program_prefetch}
return %0 : tuple<tensor<2x3xi32>, tensor<i32>>
}
func.func @main(%arg0 : tuple<tensor<2x3xi32>, tensor<i32>>) -> tuple<tensor<2x3xi32>, tensor<i32>> {
%1 = "mhlo.async_start"(%arg0) {called_computation=@copy}
%2 = "mhlo.async_done"(%1) {called_computation=@copy}
return %2 : tuple<tensor<2x3xi32>, tensor<i32>>
}
}
The parameter = 0
tells us that the async copy of the 0
th parameter is a cross_program_prefetch
, while the index
of [0]
tells us that the 0
th element of the tuple is prefetched while the other element of the tuple is not.
Parameter:
Parameter | tipe C++ | Keterangan |
---|---|---|
parameter | int64_t | |
indeks | ::llvm::ArrayRef<int64_t> | Dimensi |
mengimbangi | std::optional<int64_t> |
CustomCallScheduleAttr
Specifies the desired schedule for the custom-call.
Sintaksis:
#mhlo.custom_call_schedule<
::mlir::mhlo::CustomCallSchedule # value
>
Kasus enum:
- TIDAK ADA (
NONE
) - LATEST (
LATEST
) - EARLIEST (
EARLIEST
) #### Parameters:
Parameter | tipe C++ | Keterangan |
---|---|---|
nilai | ::mlir::mhlo::CustomCallSchedule | an enum of type CustomCallSchedule |
DequantizeModeAttr
Dequantization mode. Only MIN_COMBINED is supported.
Sintaksis:
#mhlo.dequantize_mode<
::mlir::mhlo::DequantizeMode # value
>
Kasus enum:
- MIN_COMBINED (
MIN_COMBINED
) #### Parameters:
Parameter | tipe C++ | Keterangan |
---|---|---|
nilai | ::mlir::mhlo::DequantizeMode | an enum of type DequantizeMode |
DomainKindAttr
Kind of domain metatdata attached to an HLO domain.
Sintaksis:
#mhlo.kind<
::mlir::mhlo::DomainKind # value
>
Kasus enum:
- sharding (
sharding
) #### Parameters:
Parameter | tipe C++ | Keterangan |
---|---|---|
nilai | ::mlir::mhlo::DomainKind | an enum of type DomainKind |
DotDimensionNumbersAttr
Attribute that models the dimension information for dot.
Parameter:
Parameter | tipe C++ | Keterangan |
---|---|---|
lhsBatchingDimensions | ::llvm::ArrayRef<int64_t> | Dimensi |
rhsBatchingDimensions | ::llvm::ArrayRef<int64_t> | Dimensi |
lhsContractingDimensions | ::llvm::ArrayRef<int64_t> | Dimensi |
rhsContractingDimensions | ::llvm::ArrayRef<int64_t> | Dimensi |
FftTypeAttr
XLA fast fourier transform type.
Sintaksis:
#mhlo.fft_type<
::mlir::mhlo::FftType # value
>
Kasus enum:
- FFT (
FFT
) - IFFT (
IFFT
) - RFFT (
RFFT
) - IRFFT (
IRFFT
) #### Parameters:
Parameter | tipe C++ | Keterangan |
---|---|---|
nilai | ::mlir::mhlo::FftType | an enum of type FftType |
FusionKindAttr
fusion kind
Sintaksis:
#mhlo.fusion_kind<
::mlir::mhlo::FusionKind # value
>
Kasus enum:
- kLoop (
kLoop
) - kInput (
kInput
) - kOutput (
kOutput
) - kCustom (
kCustom
) #### Parameters:
Parameter | tipe C++ | Keterangan |
---|---|---|
nilai | ::mlir::mhlo::FusionKind | an enum of type FusionKind |
GatherDimensionNumbersAttr
Attribute that models the dimension information for gather
Parameter:
Parameter | tipe C++ | Keterangan |
---|---|---|
offsetDims | ::llvm::ArrayRef<int64_t> | Dimensi |
collapsedSliceDims | ::llvm::ArrayRef<int64_t> | Dimensi |
startIndexMap | ::llvm::ArrayRef<int64_t> | Dimensi |
indexVectorDim | int64_t |
OutputOperandAliasAttr
Attribute that models the alias relationship of output and operand of a CustomCall op
Sintaksis:
#mhlo.output_operand_alias<
::llvm::ArrayRef<int64_t>, # outputTupleIndices
int64_t, # operandIndex
::llvm::ArrayRef<int64_t> # operandTupleIndices
>
This attribute captures the alias relationship of the output to one of the operands for a CustomCall op, denoted by operand_index
. The output_tuple_indices
and operand_tuple_indices
are used to index into output and operand types. These indices lists are empty if the corresponding types are not tuple types, and can be arbitrarily long in case of arbitrarily nested tuple types.
See https://www.tensorflow.org/xla/aliasing
Example when used as array with in mhlo.custom-call:
%0 = "mhlo.custom_call"(%arg0, %arg1) {
// other attributes
output_operand_alias = [
#mhlo.output_operand_alias<output_tuple_indices = [0],
operand_index = 0,
operand_tuple_indices = [1]>
]
} : (tuple<tensor<1x1xf32>, tensor<2x3xf32>>, tensor<5x5xf32>) -> tuple<tensor<2x3xf32>>
The output and the 0th operand are both tuples. The aliasing shows the
relationship between the 0th element in output tuple with the 1st element in
the 0th operand. And both of them are of the same type: tensor<2x3xf32>.
Parameter:
Parameter | tipe C++ | Keterangan |
---|---|---|
outputTupleIndices | ::llvm::ArrayRef<int64_t> | Dimensi |
operandIndex | int64_t | |
operandTupleIndices | ::llvm::ArrayRef<int64_t> | Dimensi |
PrecisionAttr
XLA precision for an operand. Has backend specific meaning.
Sintaksis:
#mhlo.precision<
::mlir::mhlo::Precision # value
>
Kasus enum:
- DEFAULT (
DEFAULT
) - HIGH (
HIGH
) - HIGHEST (
HIGHEST
) - PACKED_NIBBLE (
PACKED_NIBBLE
) #### Parameters:
Parameter | tipe C++ | Keterangan |
---|---|---|
nilai | ::mlir::mhlo::Precision | an enum of type Precision |
RngAlgorithmAttr
XLA PRNG algorithm to be used.
Sintaksis:
#mhlo.rng_algorithm<
::mlir::mhlo::RngAlgorithm # value
>
Kasus enum:
- DEFAULT (
DEFAULT
) - THREE_FRY (
THREE_FRY
) - PHILOX (
PHILOX
) #### Parameters:
Parameter | tipe C++ | Keterangan |
---|---|---|
nilai | ::mlir::mhlo::RngAlgorithm | an enum of type RngAlgorithm |
RngDistributionAttr
XLA PRNG distribution to be used.
Sintaksis:
#mhlo.rng_distribution<
::mlir::mhlo::RngDistribution # value
>
Kasus enum:
- UNIFORM (
UNIFORM
) - NORMAL (
NORMAL
) #### Parameters:
Parameter | tipe C++ | Keterangan |
---|---|---|
nilai | ::mlir::mhlo::RngDistribution | an enum of type RngDistribution |
ScatterDimensionNumbersAttr
Attribute that models the dimension information for scatter
Parameter:
Parameter | tipe C++ | Keterangan |
---|---|---|
updateWindowDims | ::llvm::ArrayRef<int64_t> | Dimensi |
insertedWindowDims | ::llvm::ArrayRef<int64_t> | Dimensi |
scatterDimsToOperandDims | ::llvm::ArrayRef<int64_t> | Dimensi |
indexVectorDim | int64_t |
SparsityDescriptorAttr
Describes structured (N:M) sparsity configuration
Sintaksis:
#mhlo.sparsity<
int64_t, # dimension
int64_t, # n
int64_t # m
>
This attribute is defined for a sparse dot operation with a structured sparse input tensor. With (N=2,M=4), every 4 consecutive logical elements have exactly 2 non-zero physical elements in the input tensor.
$dimension defines the index of the contracting dimension that is sparse (it has to be the most minor dimension). The additional metadata operand in the sparse dot operation defines which logical elements are zeroed out.
Parameter:
Parameter | tipe C++ | Keterangan |
---|---|---|
dimensi | int64_t | |
N | int64_t | |
M | int64_t |
TransposeAttr
Transpose options
Sintaksis:
#mhlo.transpose<
::mlir::mhlo::Transpose # value
>
Kasus enum:
- TRANSPOSE_INVALID (
TRANSPOSE_INVALID
) - NO_TRANSPOSE (
NO_TRANSPOSE
) - TRANSPOSE (
TRANSPOSE
) - ADJOINT (
ADJOINT
) #### Parameters:
Parameter | tipe C++ | Keterangan |
---|---|---|
nilai | ::mlir::mhlo::Transpose | an enum of type Transpose |
TypeExtensionsAttr
Attribute that extends tensor type with MHLO type properties.
Sintaksis:
#mhlo.type_extensions<
::llvm::ArrayRef<int64_t> # bounds
>
This attribute is used to extend MLIR tensor type with MHLO tensor specific properties. These properties aren't modeled in the MLIR type. This attribute is set in the encoding
field of the tensor type.
See HLO_BoundedAttrInterface
for documentation for bounds
.
Parameter:
Parameter | tipe C++ | Keterangan |
---|---|---|
batas | ::llvm::ArrayRef<int64_t> |
Tipe
AsyncBundleType
Opaque collection of other types
Sintaksis:
!mhlo.async_bundle<
::llvm::ArrayRef<Type> # types
>
Parameter:
Parameter | tipe C++ | Keterangan |
---|---|---|
jenis | ::llvm::ArrayRef<Type> |