ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-20 14:59:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

Regulate reduction accumulator operations and Element-wise operations (#274)

Browse Source 
* Remove template from Reducton operation classes and add template to their operator() and GetIdentityValue() interfaces

* Change to unary elementwise operators and the reduce_unary_operator (class for mapping) and dependent variations in all host layers

* Remove the data type template parameter from reduce_binary_operator (class for mapping) and dependent variations in host layers

* Add InMemoryDataOperatonSupportedOnDataType to check the matching between data type and InMemoryDataOperation

* Use struct-scope operator template instantiation for binary and unary element-wise operations

* Change a few more elementwise operations to use template for operator()

* Tiny correction in Normalize operator

* Add static_assert to check the data type appliability for some reduction accumulator and element-wise operatons

* Correction in some examples with regard to using ReduceAccDataType

* Use static_assert for UnaryDivide

* Update to merged codes to use Element-wise operations and Reduction Accumulator operations correctly

* Tiny fix with regard to SetWorkSpacePointer()
This commit is contained in:
Qianfeng
2022-06-18 04:10:25 +08:00
committed by GitHub
parent 63cdd92398
commit 1f543bfa79
48 changed files with 891 additions and 837 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
example/13_pool2d_fwd/pool2d_fwd_common.hpp
View File

@@ -31,16 +31,15 @@ static void pool_host_verify(const Tensor<InDataType>& in,
const std::array<ck::index_t, 2>& in_left_pads,
const std::array<ck::index_t, 2>& /*in_right_pads*/)
{
const int32_t divider = window_spatial_lengths[0] * window_spatial_lengths[1];
const int32_t reduceLength = window_spatial_lengths[0] * window_spatial_lengths[1];
using ReduceOperation = typename ck::reduce_binary_operator<AccDataType, ReduceOpId>::opType;
using InElementwiseOperation = typename ck::
reduce_unary_operator<AccDataType, ReduceOpId, true, true>::InElementwiseOperation;
using AccElementwiseOperation = typename ck::
reduce_unary_operator<AccDataType, ReduceOpId, true, true>::AccElementwiseOperation;
using ReduceOperation = typename ck::reduce_binary_operator<ReduceOpId>::opType;
const InElementwiseOperation in_elementwise_op(divider);
const AccElementwiseOperation acc_elementwise_op(divider);
auto elementwise_ops =
ck::reduce_unary_operator<ReduceOpId, true, true>::GetElementwiseOperator(reduceLength);
auto in_elementwise_op = std::get<0>(elementwise_ops);
auto acc_elementwise_op = std::get<1>(elementwise_ops);
if constexpr(!OutputIndex)
{
@@ -48,7 +47,7 @@ static void pool_host_verify(const Tensor<InDataType>& in,
ck::detail::AccumulateWithNanCheck<PropagateNan, ReduceOperation, AccDataType>;
auto f_nchw = [&](auto n, auto c, auto ho, auto wo) {
auto accuVal = ReduceOperation::GetIdentityValue();
auto accuVal = ReduceOperation::template GetIdentityValue<AccDataType>();
for(ck::index_t y = 0; y < window_spatial_lengths[0]; ++y)
{
@@ -86,7 +85,7 @@ static void pool_host_verify(const Tensor<InDataType>& in,
AccDataType,
IndexDataType>;
auto f_nchw = [&](auto n, auto c, auto ho, auto wo) {
auto accuVal = ReduceOperation::GetIdentityValue();
auto accuVal = ReduceOperation::template GetIdentityValue<AccDataType>();
IndexDataType accuIndex = 0;
for(ck::index_t y = 0; y < window_spatial_lengths[0]; ++y)