Improve normalization (#580)

* Sync the order of type string with template parameter

* Add more instances

* Check the vector size and remove redundant var

* Extract var to static, prepare to separate sweep once kernel

* Separate sweeponce flow and optimize the flow

* 1. Rename AccDatatype in normalization to computeData
2. Rename AccElementwiseOperation to YElementwiseOperation in normalization

* Remove useless code

* Update naive variance kernel

* Refine string

* Fix typo

* Support naive variance for device_normalization

* Check the blocksize

* Share the VGPR of x and y

* Share the VGPR of gamma and beta

* Add more instances

* Support fp16 sqrt for experiment

* Add CHANGELOG

* Fix typo

* clang-format

[ROCm/composable_kernel commit: 6a6163a3d1]
This commit is contained in:
rocking5566
2023-02-16 01:59:35 +08:00
committed by GitHub
parent d6de9bdcbe
commit d5062679f1
17 changed files with 826 additions and 459 deletions

View File

@@ -12,12 +12,12 @@
#include "ck/library/tensor_operation_instance/gpu/normalization.hpp"
using XDataType = ck::half_t;
using GammaDataType = ck::half_t;
using BetaDataType = ck::half_t;
using YDataType = ck::half_t;
using AccDataType = float;
using PassThrough = ck::tensor_operation::element_wise::PassThrough;
using XDataType = ck::half_t;
using GammaDataType = ck::half_t;
using BetaDataType = ck::half_t;
using YDataType = ck::half_t;
using ComputeDataType = float;
using PassThrough = ck::tensor_operation::element_wise::PassThrough;
constexpr int Rank = 2;
constexpr int NumReduceDim = 1;
@@ -54,7 +54,7 @@ int main(int argc, char* argv[])
using DeviceOp = ck::tensor_operation::device::DeviceNormalization<XDataType,
GammaDataType,
BetaDataType,
AccDataType,
ComputeDataType,
YDataType,
PassThrough,
Rank,