/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck/tensor_operation/gpu/grid/batchnorm_multiblock/gridwise_multiblock_batchnorm_forward.hpp File Reference

• develop
• include
• ck
• tensor_operation
• gpu
• grid
• batchnorm_multiblock

Classes | Namespaces | Functions

gridwise_multiblock_batchnorm_forward.hpp File Reference

#include "ck/utility/data_type.hpp"
#include "ck/utility/math_v2.hpp"
#include "ck/tensor_operation/gpu/block/blockwise_welford.hpp"
#include "ck/tensor_operation/gpu/thread/threadwise_welford.hpp"
#include "ck/tensor_operation/gpu/thread/threadwise_tensor_slice_transfer.hpp"
#include "ck/tensor_operation/gpu/element/element_wise_operation.hpp"
#include "ck/utility/workgroup_synchronization.hpp"

Go to the source code of this file.

Classes
struct	ck::GridwiseMultiblockBatchNormForward< XDataType, YDataType, AccDataType, ScaleDataType, BiasDataType, MeanVarDataType, YElementwiseOp, XYGridDesc_M_K, MeanVarCountGridDesc_M_G, MeanVarCountGridDesc_M_K, ScaleBiasGridDesc_M, MeanVarGridDesc_M, GetReduceCountPerThreadFunctor, BlockSize, MThreadClusterSize, KThreadClusterSize, MThreadSliceSize, KThreadSliceSize, XSrcYDstVectorDim, XSrcVectorSize, YDstVectorSize, ScaleSrcVectorSize, BiasSrcVectorSize, MeanVarSrcDstVectorSize >

Namespaces
	ck

Functions
template<typename GridwiseMultiblockBatchNormForward_ , typename XDataType , typename YDataType , typename AccDataType , typename ScaleDataType , typename BiasDataType , typename MeanVarDataType , typename YElementwiseOp , typename XYGridDesc_M_K , typename MeanVarCountGridDesc_M_G , typename MeanVarCountGridDesc_M_K , typename ScaleBiasGridDesc_M , typename MeanVarGridDesc_M , typename GetReduceCountPerThreadFunctor >
__global__ void	ck::kernel_multiblock_batchnorm_forward (const XYGridDesc_M_K x_grid_desc_m_k, const XYGridDesc_M_K y_grid_desc_m_k, const MeanVarCountGridDesc_M_G mean_var_count_grid_desc_m_g, const MeanVarCountGridDesc_M_K mean_var_count_grid_desc_m_k, const ScaleBiasGridDesc_M scale_grid_desc_m, const ScaleBiasGridDesc_M bias_grid_desc_m, const MeanVarGridDesc_M mean_var_grid_desc_m, const GetReduceCountPerThreadFunctor get_reduce_count_per_thread, index_t num_k_block_tile_iteration, AccDataType epsilon, const XDataType *const __restrict__ p_x, MeanVarDataType *const __restrict__ p_welford_mean, MeanVarDataType *const __restrict__ p_welford_variance, int32_t *const __restrict__ p_welford_count, int32_t *const __restrict__ p_control, const ScaleDataType *const __restrict__ p_scale, const BiasDataType *const __restrict__ p_bias, const YElementwiseOp y_elementwise_op, YDataType *const __restrict__ p_y, bool updateMovingAverage, AccDataType averageFactor, MeanVarDataType *const __restrict__ resultRunningMean, MeanVarDataType *const __restrict__ resultRunningVariance, bool saveMeanInvVariance, MeanVarDataType *const __restrict__ resultSaveMean, MeanVarDataType *const __restrict__ resultSaveInvVariance)