ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-04 13:41:24 +00:00
Code Issues Packages Projects Releases Wiki Activity

[CK_TILE] Fix compiler related FA bwd issues (#1530)

Browse Source 
* add barriers

* tail bias barriers

* adjust bf16/hd256 tol

* continue adjust bf16/hd256 tol
This commit is contained in:
Dan Yao
2024-09-27 03:18:39 +08:00
committed by GitHub
parent 42e6dceacc
commit 9d69a099a4
2 changed files with 21 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
example/ck_tile/01_fmha/fmha_bwd.cpp
View File

@@ -99,13 +99,26 @@ auto create_args(int argc, char* argv[])
// different threshold for different dtype
template <typename DataType>
auto get_elimit(int /*init_method*/)
auto get_elimit(ck_tile::index_t /*hdim_q*/, ck_tile::index_t /*hdim_v*/)
{
double rtol = 1e-2;
double atol = 1e-2;
return ck_tile::make_tuple(rtol, atol);
}
template <>
auto get_elimit<ck_tile::bf16_t>(ck_tile::index_t hdim_q, ck_tile::index_t hdim_v)
{
double rtol = 1e-2;
double atol = 1e-2;
if(hdim_q > 128 && hdim_v > 128) // 3.2 for RTZ/1.5 for RTN
{
rtol = 3.2e-2;
atol = 3.2e-2;
}
return ck_tile::make_tuple(rtol, atol);
}
template <typename DataType>
bool run(const ck_tile::ArgParser& arg_parser)
{
@@ -899,7 +912,7 @@ bool run(const ck_tile::ArgParser& arg_parser)
}
// clang-format on
auto [rtol, atol] = get_elimit<DataType>(init_method);
auto [rtol, atol] = get_elimit<DataType>(hdim_q, hdim_v);
bool dq_cur_pass = ck_tile::check_err(dq_host_result,
dq_host_ref,
std::string("Error: QGrad Incorrect results!"),