Merge branch 'develop' of...

Merge branch 'develop' of https://github.com/ROCmSoftwarePlatform/composable_kernel into navi3x_mD_batchedGEMM_GroupConvFwd

Merge branch 'develop' of...
Merge branch 'develop' of https://github.com/ROCmSoftwarePlatform/composable_kernel into navi3x_mD_batchedGEMM_GroupConvFwd
f1b53d78 · aska-0096 · 0c9cdbce · 7494c1c6 · f1b53d78 · f1b53d78
Commit f1b53d78 authored Jan 30, 2023 by aska-0096
20 changed files
--- a/include/ck/tensor_operation/gpu/element/element_wise_operation.hpp
+++ b/include/ck/tensor_operation/gpu/element/element_wise_operation.hpp
--- a/include/ck/tensor_operation/gpu/grid/block_to_ctile_map.hpp
+++ b/include/ck/tensor_operation/gpu/grid/block_to_ctile_map.hpp
--- a/include/ck/tensor_operation/gpu/grid/gemm_layernorm/gridwise_gemm_multiple_d_welford_first_half_xdl_cshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/grid/gemm_layernorm/gridwise_gemm_multiple_d_welford_first_half_xdl_cshuffle.hpp
--- a/include/ck/tensor_operation/gpu/grid/gemm_layernorm/gridwise_welford_second_half_layernorm2d.hpp
+++ b/include/ck/tensor_operation/gpu/grid/gemm_layernorm/gridwise_welford_second_half_layernorm2d.hpp
--- a/include/ck/tensor_operation/gpu/grid/gridwise_elementwise_layernorm_welford_variance.hpp
+++ b/include/ck/tensor_operation/gpu/grid/gridwise_elementwise_layernorm_welford_variance.hpp
@@ -434,7 +434,7 @@ struct GridwiseElementwiseLayernormWelfordVariance_mk_to_mk
            });

            static_for<0, MThreadSliceSize, 1>{}([&](auto iM) {
-                auto divisor = 1 / __builtin_amdgcn_sqrtf(var_thread_buf(iM) + epsilon);
+                auto divisor = 1 / ck::math::sqrt(var_thread_buf(iM) + epsilon);
                static_for<0, XThreadBufferNumber, 1>{}([&](auto iK0) {
                    static_for<0, XSrcVectorSize, 1>{}([&](auto iK1) {
                        constexpr auto offset_m_k =

--- a/include/ck/tensor_operation/gpu/grid/gridwise_gemm_waveletmodel.hpp
+++ b/include/ck/tensor_operation/gpu/grid/gridwise_gemm_waveletmodel.hpp
--- a/include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdl_waveletmodel_cshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdl_waveletmodel_cshuffle.hpp
--- a/include/ck/tensor_operation/gpu/grid/gridwise_normalization_welford_variance.hpp
+++ b/include/ck/tensor_operation/gpu/grid/gridwise_normalization_welford_variance.hpp
@@ -319,7 +319,7 @@ struct GridwiseNormalizationWelfordVariance_mk_to_mk
            });

            static_for<0, MThreadSliceSize, 1>{}([&](auto iM) {
-                auto divisor = 1 / __builtin_amdgcn_sqrtf(var_thread_buf(iM) + epsilon);
+                auto divisor = 1 / ck::math::sqrt(var_thread_buf(iM) + epsilon);
                static_for<0, XThreadBufferNumber, 1>{}([&](auto iK0) {
                    static_for<0, XSrcVectorSize, 1>{}([&](auto iK1) {
                        constexpr auto offset_m_k =

--- a/include/ck/tensor_operation/gpu/grid/gridwise_sparse_embedding3_forward_layernorm.hpp
+++ b/include/ck/tensor_operation/gpu/grid/gridwise_sparse_embedding3_forward_layernorm.hpp
--- a/include/ck/utility/reduction_operator.hpp
+++ b/include/ck/utility/reduction_operator.hpp
--- a/include/ck/utility/synchronization.hpp
+++ b/include/ck/utility/synchronization.hpp
@@ -18,6 +18,7 @@ __device__ void block_sync_lds()
    __syncthreads();
 #endif
 }
+
 __device__ void s_nop()
 {
 #if 1

--- a/library/include/ck/library/reference_tensor_operation/cpu/reference_layernorm.hpp
+++ b/library/include/ck/library/reference_tensor_operation/cpu/reference_layernorm.hpp
--- a/library/include/ck/library/reference_tensor_operation/cpu/reference_reduce.hpp
+++ b/library/include/ck/library/reference_tensor_operation/cpu/reference_reduce.hpp
--- a/library/include/ck/library/reference_tensor_operation/cpu/reference_softmax.hpp
+++ b/library/include/ck/library/reference_tensor_operation/cpu/reference_softmax.hpp
--- a/library/include/ck/library/tensor_operation_instance/device_operation_instance_factory.hpp
+++ b/library/include/ck/library/tensor_operation_instance/device_operation_instance_factory.hpp
--- a/library/include/ck/library/tensor_operation_instance/gpu/batched_gemm_bias_permute.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/batched_gemm_bias_permute.hpp
--- a/library/include/ck/library/tensor_operation_instance/gpu/batchnorm_infer.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/batchnorm_infer.hpp
--- a/library/include/ck/library/tensor_operation_instance/gpu/device_elementwise_instance.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/device_elementwise_instance.hpp
--- a/library/include/ck/library/tensor_operation_instance/gpu/gemm_add_multiply.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/gemm_add_multiply.hpp
--- a/library/include/ck/library/tensor_operation_instance/gpu/reduce/device_reduce_instance_blockwise.hpp
+++ b/library/include/ck/library/tensor_operation_instance/gpu/reduce/device_reduce_instance_blockwise.hpp