ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-04 05:31:24 +00:00
Code Issues Packages Projects Releases Wiki Activity

Batchnorm inference instances, external API, client examples and gtests (#531)

Browse Source 
* File renaming and class renaming for device element-wise operation

* Add batchnorm-infer instances, external API and client example

* Add batchnorm-infer profiler module and gtests

* Remove file device_elementwise_extension.hpp and move NormalizeInInfer operation to element_wise_operation.hpp

* Remove the using of class aliasing for DeviceElementwiseForBatchNormInfer

* Rename class and file due to conflict from device_elementwise_2d.hpp

* Fix namespace in batcnnorm_infer_nhwc client example
This commit is contained in:
Qianfeng
2023-01-26 07:09:04 +08:00
committed by GitHub
parent 52abc2f371
commit a1b2441f8d
29 changed files with 1260 additions and 80 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
include/ck/tensor_operation/gpu/device/device_elementwise_base.hpp → include/ck/tensor_operation/gpu/device/device_elementwise.hpp
View File

@@ -17,7 +17,7 @@ template <typename InDataTypeTuple,
typename OutDataTypeTuple,
typename ElementwiseOperation,
index_t NumDim>
struct DeviceElementwiseBase : public BaseOperator
struct DeviceElementwise : public BaseOperator
{
static constexpr int NumInput = InDataTypeTuple::Size();
static constexpr int NumOutput = OutDataTypeTuple::Size();
@@ -37,8 +37,8 @@ template <typename InDataTypeTuple,
typename OutDataTypeTuple,
typename ElementwiseOperation,
index_t NumDim>
using DeviceElementwiseBasePtr = std::unique_ptr<
DeviceElementwiseBase<InDataTypeTuple, OutDataTypeTuple, ElementwiseOperation, NumDim>>;
using DeviceElementwisePtr = std::unique_ptr<
DeviceElementwise<InDataTypeTuple, OutDataTypeTuple, ElementwiseOperation, NumDim>>;
} // namespace device
} // namespace tensor_operation

10
include/ck/tensor_operation/gpu/device/device_elementwise_2d.hpp → include/ck/tensor_operation/gpu/device/impl/device_elementwise_2d_impl.hpp
View File

@@ -8,7 +8,7 @@
#include "ck/utility/math.hpp"
#include "ck/utility/sequence.hpp"
#include "ck/tensor_operation/gpu/device/device_elementwise_base.hpp"
#include "ck/tensor_operation/gpu/device/device_elementwise.hpp"
#include "ck/tensor_operation/gpu/grid/gridwise_elementwise_2d.hpp"
#include "ck/tensor_description/tensor_descriptor_helper.hpp"
@@ -26,10 +26,10 @@ template <typename InDataTypeTuple,
index_t NPerThread,
typename InScalarPerVectorSeq,
typename OutScalarPerVectorSeq>
struct DeviceElementwise : public DeviceElementwiseBase<InDataTypeTuple,
OutDataTypeTuple,
ElementwiseOperation,
NumDim_m + NumDim_n>
struct DeviceElementwise2dImpl : public DeviceElementwise<InDataTypeTuple,
OutDataTypeTuple,
ElementwiseOperation,
NumDim_m + NumDim_n>
{
static constexpr index_t NumDim = NumDim_m + NumDim_n;

6
include/ck/tensor_operation/gpu/device/impl/device_elementwise.hpp → include/ck/tensor_operation/gpu/device/impl/device_elementwise_impl.hpp
View File

@@ -8,7 +8,7 @@
#include "ck/utility/math.hpp"
#include "ck/utility/sequence.hpp"
#include "ck/tensor_operation/gpu/device/device_elementwise_base.hpp"
#include "ck/tensor_operation/gpu/device/device_elementwise.hpp"
#include "ck/tensor_operation/gpu/grid/gridwise_elementwise_1d.hpp"
#include "ck/tensor_description/tensor_descriptor_helper.hpp"
@@ -25,8 +25,8 @@ template <typename InDataTypeTuple,
index_t MPerThread,
typename InScalarPerVectorSeq,
typename OutScalarPerVectorSeq>
struct DeviceElementwise
: public DeviceElementwiseBase<InDataTypeTuple, OutDataTypeTuple, ElementwiseOperation, NumDim>
struct DeviceElementwiseImpl
: public DeviceElementwise<InDataTypeTuple, OutDataTypeTuple, ElementwiseOperation, NumDim>
{
static constexpr int NumInput = InDataTypeTuple::Size();
static constexpr int NumOutput = OutDataTypeTuple::Size();