Skip to content

GitLab

Commit a30c626b authored by Bartlomiej Wroblewski's avatar Bartlomiej Wroblewski
Browse files

Make ComputeDataType an optional argument

parent b019d839
Expand all Show whitespace changes
Inline Side-by-side
Showing with 148 additions and 156 deletions
client_example/04_contraction/contraction_bilinear_fp32.cpp
View file @ a30c626b
...@@ -30,7 +30,6 @@ using CShuffleDataType = F32; ...@@ -30,7 +30,6 @@ using CShuffleDataType = F32;
using DDataType = F32; using DDataType = F32;
using DsDataType = ck::Tuple<DDataType>; using DsDataType = ck::Tuple<DDataType>;
using EDataType = F32; using EDataType = F32;
using ComputeDataType = F32;
static constexpr ck::index_t NumDimM = 2; static constexpr ck::index_t NumDimM = 2;
static constexpr ck::index_t NumDimN = 2; static constexpr ck::index_t NumDimN = 2;
...@@ -141,7 +140,6 @@ int main(int argc, char* argv[]) ...@@ -141,7 +140,6 @@ int main(int argc, char* argv[])
BDataType, BDataType,
ck::Tuple<DDataType>, ck::Tuple<DDataType>,
EDataType, EDataType,
ComputeDataType,
ck::tensor_operation::element_wise::PassThrough, ck::tensor_operation::element_wise::PassThrough,
ck::tensor_operation::element_wise::PassThrough, ck::tensor_operation::element_wise::PassThrough,
ck::tensor_operation::element_wise::Bilinear>; ck::tensor_operation::element_wise::Bilinear>;
......
client_example/04_contraction/contraction_bilinear_fp64.cpp
View file @ a30c626b
...@@ -30,7 +30,6 @@ using CShuffleDataType = F64; ...@@ -30,7 +30,6 @@ using CShuffleDataType = F64;
using DDataType = F64; using DDataType = F64;
using DsDataType = ck::Tuple<DDataType>; using DsDataType = ck::Tuple<DDataType>;
using EDataType = F64; using EDataType = F64;
using ComputeDataType = F64;
static constexpr ck::index_t NumDimM = 2; static constexpr ck::index_t NumDimM = 2;
static constexpr ck::index_t NumDimN = 2; static constexpr ck::index_t NumDimN = 2;
...@@ -186,7 +185,6 @@ int main(int argc, char* argv[]) ...@@ -186,7 +185,6 @@ int main(int argc, char* argv[])
BDataType, BDataType,
ck::Tuple<DDataType>, ck::Tuple<DDataType>,
EDataType, EDataType,
ComputeDataType,
ck::tensor_operation::element_wise::PassThrough, ck::tensor_operation::element_wise::PassThrough,
ck::tensor_operation::element_wise::PassThrough, ck::tensor_operation::element_wise::PassThrough,
ck::tensor_operation::element_wise::Bilinear>; ck::tensor_operation::element_wise::Bilinear>;
......
client_example/04_contraction/contraction_scale_fp32.cpp
View file @ a30c626b
...@@ -29,7 +29,6 @@ using AccDataType = F32; ...@@ -29,7 +29,6 @@ using AccDataType = F32;
using CShuffleDataType = F32; using CShuffleDataType = F32;
using DsDataType = ck::Tuple<>; using DsDataType = ck::Tuple<>;
using EDataType = F32; using EDataType = F32;
using ComputeDataType = F32;
static constexpr ck::index_t NumDimM = 2; static constexpr ck::index_t NumDimM = 2;
static constexpr ck::index_t NumDimN = 2; static constexpr ck::index_t NumDimN = 2;
...@@ -128,7 +127,6 @@ int main(int argc, char* argv[]) ...@@ -128,7 +127,6 @@ int main(int argc, char* argv[])
BDataType, BDataType,
ck::Tuple<>, ck::Tuple<>,
EDataType, EDataType,
ComputeDataType,
ck::tensor_operation::element_wise::PassThrough, ck::tensor_operation::element_wise::PassThrough,
ck::tensor_operation::element_wise::PassThrough, ck::tensor_operation::element_wise::PassThrough,
ck::tensor_operation::element_wise::Scale>; ck::tensor_operation::element_wise::Scale>;
......
client_example/04_contraction/contraction_scale_fp64.cpp
View file @ a30c626b
...@@ -29,7 +29,6 @@ using AccDataType = F64; ...@@ -29,7 +29,6 @@ using AccDataType = F64;
using CShuffleDataType = F64; using CShuffleDataType = F64;
using DsDataType = ck::Tuple<>; using DsDataType = ck::Tuple<>;
using EDataType = F64; using EDataType = F64;
using ComputeDataType = F64;
static constexpr ck::index_t NumDimM = 2; static constexpr ck::index_t NumDimM = 2;
static constexpr ck::index_t NumDimN = 2; static constexpr ck::index_t NumDimN = 2;
...@@ -176,7 +175,6 @@ int main(int argc, char* argv[]) ...@@ -176,7 +175,6 @@ int main(int argc, char* argv[])
BDataType, BDataType,
ck::Tuple<>, ck::Tuple<>,
EDataType, EDataType,
ComputeDataType,
ck::tensor_operation::element_wise::PassThrough, ck::tensor_operation::element_wise::PassThrough,
ck::tensor_operation::element_wise::PassThrough, ck::tensor_operation::element_wise::PassThrough,
ck::tensor_operation::element_wise::Scale>; ck::tensor_operation::element_wise::Scale>;
......
include/ck/tensor_operation/gpu/device/device_contraction_multiple_d.hpp
View file @ a30c626b
...@@ -31,10 +31,10 @@ template <index_t NumDimM, ...@@ -31,10 +31,10 @@ template <index_t NumDimM,
typename BDataType, typename BDataType,
typename DsDataType, typename DsDataType,
typename EDataType, typename EDataType,
typename ComputeDataType,
typename AElementwiseOperation, typename AElementwiseOperation,
typename BElementwiseOperation, typename BElementwiseOperation,
typename CDEElementwiseOperation> typename CDEElementwiseOperation,
typename ComputeDataType = ADataType>
struct DeviceContractionMultipleD : public BaseOperator struct DeviceContractionMultipleD : public BaseOperator
{ {
static constexpr index_t NumDTensor = DsDataType::Size(); static constexpr index_t NumDTensor = DsDataType::Size();
......
include/ck/tensor_operation/gpu/device/impl/device_contraction_multiple_d_xdl_cshuffle.hpp
View file @ a30c626b
...@@ -155,10 +155,10 @@ struct DeviceContractionMultipleD_Xdl_CShuffle ...@@ -155,10 +155,10 @@ struct DeviceContractionMultipleD_Xdl_CShuffle
BDataType, BDataType,
DsDataType, DsDataType,
EDataType, EDataType,
ComputeDataType,
AElementwiseOperation, AElementwiseOperation,
BElementwiseOperation, BElementwiseOperation,
CDEElementwiseOperation> CDEElementwiseOperation,
ComputeDataType>
{ {
using DeviceOp = DeviceContractionMultipleD_Xdl_CShuffle; using DeviceOp = DeviceContractionMultipleD_Xdl_CShuffle;
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_compute_f32_kknn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_ ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_
BF16, BF16,
BF16_Tuple, BF16_Tuple,
BF16, BF16,
F32,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
F32>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_compute_f32_knnn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_ ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_
BF16, BF16,
BF16_Tuple, BF16_Tuple,
BF16, BF16,
F32,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
F32>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_compute_f32_mknn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_ ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_
BF16, BF16,
BF16_Tuple, BF16_Tuple,
BF16, BF16,
F32,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
F32>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_compute_f32_mnnn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_ ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_bf16_bf16_bf16_bf16_
BF16, BF16,
BF16_Tuple, BF16_Tuple,
BF16, BF16,
F32,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
F32>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_compute_f32_kknn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_comp ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_comp
F16, F16,
F16_Tuple, F16_Tuple,
F16, F16,
F32,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
F32>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_compute_f32_knnn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_comp ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_comp
F16, F16,
F16_Tuple, F16_Tuple,
F16, F16,
F32,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
F32>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_compute_f32_mknn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_comp ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_comp
F16, F16,
F16_Tuple, F16_Tuple,
F16, F16,
F32,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
F32>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_compute_f32_mnnn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_comp ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f16_f16_f16_f16_comp
F16, F16,
F16_Tuple, F16_Tuple,
F16, F16,
F32,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
F32>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_compute_bf16_kknn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_comp ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_comp
F32, F32,
F32_Tuple, F32_Tuple,
F32, F32,
BF16,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
BF16>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_compute_bf16_knnn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_comp ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_comp
F32, F32,
F32_Tuple, F32_Tuple,
F32, F32,
BF16,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
BF16>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_compute_bf16_mknn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_comp ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_comp
F32, F32,
F32_Tuple, F32_Tuple,
F32, F32,
BF16,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
BF16>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
library/src/tensor_operation_instance/gpu/contraction_bilinear/device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_compute_bf16_mnnn_instance.cpp
View file @ a30c626b
...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_comp ...@@ -41,10 +41,10 @@ void add_device_contraction_bilinear_m2_n2_k2_xdl_c_shuffle_f32_f32_f32_f32_comp
F32, F32,
F32_Tuple, F32_Tuple,
F32, F32,
BF16,
PassThrough, PassThrough,
PassThrough, PassThrough,
Bilinear>>>& instances) Bilinear,
BF16>>>& instances)
{ {
add_device_operation_instances( add_device_operation_instances(
instances, instances,
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment