ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-18 03:49:41 +00:00
Code Issues Packages Projects Releases Wiki Activity

Extend XDL kernel to Support RDNA3/4 - Part 5 (#2725)

Browse Source 
* Enable xdl in gfx11 & gfx12

* update cmake file

* fix all instance build (cmake)

* fix batched_gemm_gemm(cmake)

* rebase cmake files

* fix cmake build error

* remve CK_ENABLE_DYNAMIC_WARP_SIZE

* update cmake build error2

* fix gfx11 build

CK_USE_XDL is enabled on gfx11 and gfx12

* fix gfx10 build

* fix gfx11 error

---------

Co-authored-by: Lin, Qun <Quentin.Lin+amdeng@amd.com>

[ROCm/composable_kernel commit: f22740df82]
This commit is contained in:
linqunAMD
2025-09-16 01:59:25 +08:00
committed by GitHub
parent 8cbf571d53
commit a9e6cb0ec0
33 changed files with 243 additions and 397 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
include/ck/host_utility/device_prop.hpp
View File

@@ -68,11 +68,8 @@ inline bool is_gfx11_supported()
inline bool is_xdl_supported()
{
return ck::get_device_name() == "gfx908" || ck::get_device_name() == "gfx90a" ||
ck::get_device_name() == "gfx942" || ck::get_device_name() == "gfx950"
#if defined(CK_ENABLE_DYNAMIC_WARP_SIZE)
|| is_gfx12_supported() || is_gfx11_supported()
#endif
;
ck::get_device_name() == "gfx942" || ck::get_device_name() == "gfx950" ||
is_gfx12_supported() || is_gfx11_supported();
}
template <typename ADataType, typename BDataType, index_t MPerXDL, index_t NPerXDL>
@@ -83,7 +80,6 @@ inline bool is_xdl_wmma_supported()
{
return true;
}
#if defined(CK_ENABLE_DYNAMIC_WARP_SIZE)
else if(is_gfx12_supported() || is_gfx11_supported())
{
if constexpr((MPerXDL != 16) || (NPerXDL != 16))
@@ -96,7 +92,6 @@ inline bool is_xdl_wmma_supported()
}
return true;
}
#endif
else
{
return false;

11
include/ck/utility/get_id.hpp
View File

@@ -7,7 +7,6 @@
namespace ck {
#if defined(CK_ENABLE_DYNAMIC_WARP_SIZE)
__device__ constexpr index_t get_warp_size()
{
#if defined(__HIP_DEVICE_COMPILE__)
@@ -38,16 +37,6 @@ inline __host__ index_t get_warp_size()
#endif
return 64;
}
#else
__host__ __device__ constexpr index_t get_warp_size()
{
#if defined(__GFX9__) || !defined(__HIP_DEVICE_COMPILE__)
return 64;
#else
return 32;
#endif
}
#endif
__device__ index_t get_thread_local_1d_id() { return threadIdx.x; }