leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 25 Packages Projects Releases Wiki Activity

v4.2 release. (#2587)

Browse Source 
* Fix default cluster callback values to 1 to avoid profiler failure when these values are not set in command line.

* v4.2 release.
This commit is contained in:
Junkai-Wu
2025-08-22 18:11:24 -04:00
committed by GitHub
parent 11cad1f67b
commit a49a78ffef
351 changed files with 28182 additions and 2032 deletions
Download Patch File Download Diff File Expand all files Collapse all files
customConfigs.cmake
+6 -1
View File
@@ -65,7 +65,12 @@ endfunction()
if(CUTLASS_BUILD_FOR_PROFILER_REGRESSIONS)
set(PROFILER_ARCH_LIST 100a 100f 101a 101f 120a 120f)
set(PROFILER_ARCH_LIST 100a 100f 103a 120a 120f 121a)
if (CUDA_VERSION VERSION_LESS 13.0)
list(APPEND PROFILER_ARCH_LIST 101a 101f)
else()
list(APPEND PROFILER_ARCH_LIST 110a 110f)
endif()
foreach(ARCH IN LISTS CUTLASS_NVCC_ARCHS)
if(NOT (ARCH IN_LIST PROFILER_ARCH_LIST))
message(FATAL_ERROR "Only SM${PROFILER_ARCH_LIST} compute capabilities are supported with profiler-based unit tests")