Merge branch 'develop' into amd-develop

687d2b7e · Jun Liu · 5d718e6b · f5210953 · 687d2b7e · 687d2b7e
Commit 687d2b7e authored Mar 19, 2024 by Jun Liu
20 changed files
--- a/include/ck/tensor_operation/gpu/block/thread_group_tensor_slice_transfer_v4r1_dequant.hpp
+++ b/include/ck/tensor_operation/gpu/block/thread_group_tensor_slice_transfer_v4r1_dequant.hpp
--- a/include/ck/tensor_operation/gpu/device/device_gemm_dequantB.hpp
+++ b/include/ck/tensor_operation/gpu/device/device_gemm_dequantB.hpp
--- a/include/ck/tensor_operation/gpu/device/impl/device_batched_contraction_multiple_d_wmma_cshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_batched_contraction_multiple_d_wmma_cshuffle.hpp
--- a/include/ck/tensor_operation/gpu/device/impl/device_batched_gemm_softmax_gemm_permute_wmma_cshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_batched_gemm_softmax_gemm_permute_wmma_cshuffle.hpp
--- a/include/ck/tensor_operation/gpu/device/impl/device_fpAintB_gemm_wmma.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_fpAintB_gemm_wmma.hpp
--- a/include/ck/tensor_operation/gpu/device/impl/device_gemm_multiple_d_wmma_cshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_gemm_multiple_d_wmma_cshuffle.hpp
--- a/include/ck/tensor_operation/gpu/device/impl/device_gemm_multiple_d_xdl_cshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_gemm_multiple_d_xdl_cshuffle.hpp
--- a/include/ck/tensor_operation/gpu/device/impl/device_gemm_wmma.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_gemm_wmma.hpp
--- a/include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_data_multiple_d_wmma_cshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_data_multiple_d_wmma_cshuffle.hpp
--- a/include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_wmma_cshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_wmma_cshuffle.hpp
--- a/include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_multiple_d_wmma_cshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_grouped_conv_fwd_multiple_d_wmma_cshuffle.hpp
--- a/include/ck/tensor_operation/gpu/device/impl/device_grouped_query_attention_forward_wmma.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_grouped_query_attention_forward_wmma.hpp
--- a/include/ck/tensor_operation/gpu/device/impl/device_multi_query_attention_forward_wmma.hpp
+++ b/include/ck/tensor_operation/gpu/device/impl/device_multi_query_attention_forward_wmma.hpp
--- a/include/ck/tensor_operation/gpu/device/masking_specialization.hpp
+++ b/include/ck/tensor_operation/gpu/device/masking_specialization.hpp
@@ -53,7 +53,10 @@ struct MaskOutUpperTrianglePredicate
 template <typename MaskOutPredicate>
 struct C0MatrixMask_impl
 {
-    C0MatrixMask_impl(index_t NRaw) : NRaw_(NRaw), predicate_(MaskOutPredicate{}) {}
+    __host__ __device__ C0MatrixMask_impl(index_t NRaw)
+        : NRaw_(NRaw), predicate_(MaskOutPredicate{})
+    {
+    }
    __host__ __device__ constexpr bool IsNOutOfBound(/*index_t m, */ index_t n) const
    {

--- a/include/ck/tensor_operation/gpu/element/unary_element_wise_operation.hpp
+++ b/include/ck/tensor_operation/gpu/element/unary_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/gridwise_batched_gemm_multiple_d_gemm_multiple_d_xdl_cshuffle_v1.hpp
+++ b/include/ck/tensor_operation/gpu/grid/gridwise_batched_gemm_multiple_d_gemm_multiple_d_xdl_cshuffle_v1.hpp
--- a/include/ck/tensor_operation/gpu/grid/gridwise_batched_gemm_softmax_gemm_wmma_cshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/grid/gridwise_batched_gemm_softmax_gemm_wmma_cshuffle.hpp
--- a/include/ck/tensor_operation/gpu/grid/gridwise_fpAintB_gemm_wmma.hpp
+++ b/include/ck/tensor_operation/gpu/grid/gridwise_fpAintB_gemm_wmma.hpp
--- a/include/ck/tensor_operation/gpu/grid/gridwise_gemm_multiple_d_wmma_cshuffle.hpp
+++ b/include/ck/tensor_operation/gpu/grid/gridwise_gemm_multiple_d_wmma_cshuffle.hpp