msccl: add templated kernel (#945)

* msccl: add templated kernel

* Use defines to improve code readability

* Fix kernel indexing and review feedback
This commit is contained in:
Wenkai Du
2023-11-02 17:21:53 -07:00
committed by GitHub
parent 61aed56ca7
commit f484ff17b9
6 changed files with 96 additions and 73 deletions
+2 -1
View File
@@ -105,7 +105,8 @@ function(expand_collectives FILE FUNC)
#include \"primitives.h\"
#include \"collectives.h\"
#include \"devcomm.h\"
MSCCL_IMPL_KERNEL_ENTRY_FUNC_DEVREDOP_TYPE(${REDOP_CURRENT}, ${DATA_TYPE});")
MSCCL_IMPL_KERNEL_ENTRY_FUNC_DEVREDOP_TYPE(${REDOP_CURRENT}, ${DATA_TYPE}, false);
MSCCL_IMPL_KERNEL_ENTRY_FUNC_DEVREDOP_TYPE(${REDOP_CURRENT}, ${DATA_TYPE}, true);")
else()
file(WRITE ${FILE_NAME}
"#include \"${FILE}.h\"