ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-06-08 15:30:23 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix: add CK_GFX950_SUPPORT macro for gfx950 detection

Browse Source
This commit is contained in:
AviralGoelAMD
2025-07-23 14:00:32 -05:00
parent e65a716367
commit 1071198192
2 changed files with 5 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
CMakeLists.txt
View File

@@ -236,6 +236,8 @@ endif()
if (SUPPORTED_GPU_TARGETS MATCHES "gfx950")
add_definitions(-DCK_USE_NATIVE_MX_SUPPORT)
set(CK_USE_NATIVE_MX_SUPPORT "ON")
add_definitions(-DCK_GFX950_SUPPORT)
set(CK_GFX950_SUPPORT "ON")
endif()
option(CK_USE_FP8_ON_UNSUPPORTED_ARCH "Enable FP8 GEMM instances on older architectures" OFF)

6
example/ck_tile/03_gemm/gemm_utils.hpp
View File

@@ -21,7 +21,7 @@
template <typename PrecType, ck_tile::index_t M_Warp_Tile>
constexpr ck_tile::index_t get_k_warp_tile()
{
#if defined(__gfx950__)
#if defined(CK_GFX950_SUPPORT)
constexpr bool is_8bit_float =
std::is_same_v<PrecType, ck_tile::fp8_t> || std::is_same_v<PrecType, ck_tile::bf8_t>;
if constexpr(M_Warp_Tile == 32)
@@ -38,7 +38,7 @@ constexpr ck_tile::index_t get_k_warp_tile()
template <typename PrecType, ck_tile::index_t M_Warp_Tile>
constexpr ck_tile::index_t get_k_warp_tile_flatmm()
{
#if defined(__gfx950__)
#if defined(CK_GFX950_SUPPORT)
if constexpr(M_Warp_Tile == 32)
return sizeof(PrecType) == 2 ? 16 : 64;
else
@@ -293,7 +293,7 @@ struct GemmConfigPreshuffle_3 : public GemmConfigBase
static constexpr int kBlockPerCu = 2;
static constexpr auto Scheduler = ck_tile::GemmPipelineScheduler::Default;
static constexpr ck_tile::index_t Pipeline = CK_TILE_PIPELINE_PRESHUFFLE_V3;
static constexpr ck_tile::index_t Pipeline = CK_TILE_PIPELINE_PRESHUFFLE_V2;
static constexpr bool Preshuffle = true;
static constexpr bool DoubleSmemBuffer = true;
};