Fusion Conv+Bias+ReLU(+Add) (#62)

* fix relu

* clean up

* clean up

* adding 1x1 conv

* adding 1x1 conv

* added 1x1 conv

* refactor

* refactor

* refactor

* added profiler for conv+bias+relu+add

* clean up

* adding conv+bias+relu

* adding conv+bias+relu

* added conv+bias+relu

* Update README.md

* update cpu verification

* adding c shuffle

* update static_tensor for dealing with invalid element

* adding c shuffle

* debugging

* fix bug

* convert to fp16 before shuffle

* shuffle more than one M/NRepeat

* clean up

* remove coordinate step hack from GridwiseGemm_k0mk1_k0nk1_mn_xdlops_v3r1

* clean up

* remove coordinate step hack from all gridwise gemm xdl

* clean up coordinate step hack

* clean up coordinate step hack

* ThreadwiseTensorSliceTransfer_v3r2 support pointwise op on both src and dst

* adding output shuffle in conv+bias+relu+add

* update

* added conv+bias+relu+add with c shuffle

* added conv+bias+relu+add with c shuffle

* fix forward_sweep bugs in threadwise copy

* clean up

* refactor

* clean up

* clean up

* added conv_c_shuffle+bias_relu

* clean up

* added conv+bias+relu+atomic_add

* clean up

* clean up

* clean up

* clean up

* clean up

* clean up

* misc fixes; add 1x1 specialization

* clean up

* delete unused device op

* clean up

* add support for odd C value
This commit is contained in:
Chao Liu
2021-12-26 08:43:42 -06:00
committed by GitHub
parent a4f24233e5
commit acbd7bd7c5
90 changed files with 13347 additions and 2639 deletions

View File

@@ -0,0 +1,49 @@
#ifndef DEVICE_CONV_FWD_BIAS_ACTIVATION_HPP
#define DEVICE_CONV_FWD_BIAS_ACTIVATION_HPP
#include <iostream>
#include "device_base.hpp"
namespace ck {
namespace tensor_operation {
namespace device {
template <typename InElementwiseOperation,
typename WeiElementwiseOperation,
typename OutElementwiseOperation>
struct DeviceConvFwdBiasActivation : public BaseOperator
{
virtual std::unique_ptr<BaseArgument>
MakeArgumentPointer(const void* p_in,
const void* p_wei,
void* p_out,
const void* p_bias,
ck::index_t N,
ck::index_t K,
ck::index_t C,
std::vector<ck::index_t> input_spatial_lengths,
std::vector<ck::index_t> filter_spatial_lengths,
std::vector<ck::index_t> output_spatial_lengths,
std::vector<ck::index_t> conv_filter_strides,
std::vector<ck::index_t> conv_filter_dilations,
std::vector<ck::index_t> input_left_pads,
std::vector<ck::index_t> input_right_pads,
InElementwiseOperation in_element_op,
WeiElementwiseOperation wei_element_op,
OutElementwiseOperation out_element_op) = 0;
virtual std::unique_ptr<BaseInvoker> MakeInvokerPointer() = 0;
};
template <typename InElementwiseOperation,
typename WeiElementwiseOperation,
typename OutElementwiseOperation>
using DeviceConvFwdBiasActivationPtr =
std::unique_ptr<DeviceConvFwdBiasActivation<InElementwiseOperation,
WeiElementwiseOperation,
OutElementwiseOperation>>;
} // namespace device
} // namespace tensor_operation
} // namespace ck
#endif