ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-06-06 07:51:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix mismatched return type

Browse Source
This commit is contained in:
Po-Yen, Chen
2024-06-02 20:11:49 -04:00
committed by PoYen, Chen
parent 9ac2654b55
commit eac0f3cc47
2 changed files with 6 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
example/ck_tile/01_fmha/fmha_fwd.hpp
View File

@@ -431,13 +431,13 @@ template <typename Traits_>
float fmha_fwd_(const ck_tile::stream_config&, fmha_fwd_args);
template <typename Traits_>
float fmha_fwd_splitkv_oneshot_(const ck_tile::stream_config&, fmha_fwd_args);
void fmha_fwd_splitkv_oneshot_(const ck_tile::stream_config&, fmha_fwd_args);
template <typename Traits_>
std::string fmha_fwd_splitkv_get_name_();
template <typename Traits_>
float fmha_fwd_splitkv_combine_oneshot_(const ck_tile::stream_config&, fmha_fwd_args);
void fmha_fwd_splitkv_combine_oneshot_(const ck_tile::stream_config&, fmha_fwd_args);
template <typename Traits_>
std::string fmha_fwd_splitkv_combine_get_name_();

8
example/ck_tile/01_fmha/generate.py
View File

@@ -232,13 +232,13 @@ using trait_{F_idx} = fmha_fwd_traits_<{F_hdim}, {F_dtype}, {F_mode},{F_bm0}, {F
#include <iostream>
template<>
float fmha_fwd_splitkv_oneshot_<trait_{F_idx}>(const ck_tile::stream_config& s, fmha_fwd_args a)
void fmha_fwd_splitkv_oneshot_<trait_{F_idx}>(const ck_tile::stream_config& s, fmha_fwd_args a)
{{
using k_ = fmha_kernel_{F_idx};
auto [kargs, grids] = fmha_fwd_splitkv_create_kargs_and_grids<k_>(a);
constexpr dim3 blocks = k_::BlockSize();
constexpr ck_tile::index_t kBlockPerCu = k_::kBlockPerCu;
return ck_tile::launch_kernel(s, ck_tile::make_kernel<blocks.x, kBlockPerCu>(k_{{}}, grids, blocks, 0, kargs));
ck_tile::make_kernel<blocks.x, kBlockPerCu>(k_{{}}, grids, blocks, 0, kargs)(ck_tile::stream_config{{s.stream_id_}});
}}
template<>
@@ -312,13 +312,13 @@ using trait_{F_idx} = fmha_fwd_traits_<{F_hdim}, {F_dtype}, {F_mode},{F_bm0}, {F
#include <iostream>
template<>
float fmha_fwd_splitkv_combine_oneshot_<trait_{F_idx}>(const ck_tile::stream_config& s, fmha_fwd_args a)
void fmha_fwd_splitkv_combine_oneshot_<trait_{F_idx}>(const ck_tile::stream_config& s, fmha_fwd_args a)
{{
using k_ = fmha_kernel_{F_idx};
auto [kargs, grids] = fmha_fwd_splitkv_combine_create_kargs_and_grids<k_>(a);
constexpr dim3 blocks = k_::BlockSize();
constexpr ck_tile::index_t kBlockPerCu = k_::kBlockPerCu;
return ck_tile::launch_kernel(s, ck_tile::make_kernel<blocks.x, kBlockPerCu>(k_{{}}, grids, blocks, 0, kargs));
ck_tile::make_kernel<blocks.x, kBlockPerCu>(k_{{}}, grids, blocks, 0, kargs)(ck_tile::stream_config{{s.stream_id_}});
}}
template<>