mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-05 14:11:29 +00:00
Support Wave32 in CK_TILE - Part 1 (#2594)
* Support wave32/wave64 in CK_TILE - Part 1 * remove blocksize in kernel launch * fix build error * fix clang format * fix clang format 2 * fix clang format 3 * fix fmha build error * fix fmha build 2 * fix fmha build 3 * fix build error 4 * address review comment * update change log * replace KernelBlockSize with kBlockSize * fix CI fail * fix clang format * address review comment and rebase code. * fix universal test fail --------- Co-authored-by: Lin, Qun <Quentin.Lin+amdeng@amd.com> Co-authored-by: Thomas Ning <Thomas.Ning@amd.com>
This commit is contained in:
@@ -99,7 +99,6 @@ float gemm_calc_aquant(const ck_tile::AQuantGemmHostArgs& args, const ck_tile::s
|
||||
ck_tile::tuple<>,
|
||||
CLayout,
|
||||
ck_tile::element_wise::PassThrough,
|
||||
CodegenPipelineProblem::kBlockSize,
|
||||
TilePartitioner::MPerBlock,
|
||||
TilePartitioner::NPerBlock,
|
||||
M_Warp,
|
||||
@@ -114,8 +113,8 @@ float gemm_calc_aquant(const ck_tile::AQuantGemmHostArgs& args, const ck_tile::s
|
||||
|
||||
auto kargs = Kernel::MakeKernelArgs(args);
|
||||
|
||||
const dim3 grids = Kernel::GridSize(args.M, args.N, args.k_batch);
|
||||
constexpr dim3 blocks = Kernel::BlockSize();
|
||||
const dim3 grids = Kernel::GridSize(args.M, args.N, args.k_batch);
|
||||
const dim3 blocks = Kernel::BlockSize();
|
||||
|
||||
if(args.k_batch != 1)
|
||||
{
|
||||
@@ -139,7 +138,7 @@ float gemm_calc_aquant(const ck_tile::AQuantGemmHostArgs& args, const ck_tile::s
|
||||
}
|
||||
|
||||
float ave_time = ck_tile::launch_kernel(
|
||||
s, ck_tile::make_kernel<blocks.x, kBlockPerCu>(Kernel{}, grids, blocks, 0, kargs));
|
||||
s, ck_tile::make_kernel<kBlockPerCu>(Kernel{}, grids, blocks, 0, kargs));
|
||||
|
||||
return ave_time;
|
||||
};
|
||||
|
||||
Reference in New Issue
Block a user