ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-04-20 14:59:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

Set RNE fp8 conversion as a default (#1458)

Browse Source 
* Set RNE fp8 conversion as a default

* Update f8 tests

* Disable failing test on gfx11

* Update bf8 tests

* Add a flag

* Fix the flag

* Raise flag for gfx10 as well

* Temp commit for tolerance testing

* Update tolerances
This commit is contained in:
Rostyslav Geyyer
2024-08-21 11:09:48 -05:00
committed by GitHub
parent dc82daa86e
commit e20f20efbf
7 changed files with 69 additions and 50 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
include/ck/ck.hpp
View File

@@ -1,5 +1,5 @@
// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
// Copyright (c) 2018-2024, Advanced Micro Devices, Inc. All rights reserved.
#pragma once
@@ -153,8 +153,8 @@ CK_DECLARE_ENV_VAR_BOOL(CK_LOGGING)
// LDS direct loads using inline assembly
#define CK_USE_AMD_LDS_DIRECT_LOAD_INLINE_ASM 0
// set stochastic rounding as default for f8 conversions
#define CK_USE_SR_F8_CONVERSION 1
// set rounding to nearest even as default for f8 conversions
#define CK_USE_SR_F8_CONVERSION 0
// block synchronization only s_wait lgkmcnt(0), not vmcnt(0)
#define CK_EXPERIMENTAL_BLOCK_SYNC_LDS_WITHOUT_SYNC_VMEM 1