rearranged how the min and max functions are part of msccl (#1025)

* rearranged how the min and max functions are part of msccl

* added more coverage on in place graph tests
Tento commit je obsažen v:
akolliasAMD
2023-12-21 08:58:33 -07:00
odevzdal GitHub
rodič 7e1cbb440d
revize f4858e14b2
4 změnil soubory, kde provedl 8 přidání a 8 odebrání
+1 -1
Zobrazit soubor
@@ -39,8 +39,8 @@ __global__ void MSCCL_KERNEL_ENTRY_NAME(devredop, type, proto, fullOps)(struct n
#define MSCCL_DECL_KERNEL_ENTRY_FUNC() \
MSCCL_DECL_KERNEL_ENTRY_FUNC_DEVREDOP(Sum, false) \
MSCCL_DECL_KERNEL_ENTRY_FUNC_DEVREDOP(Prod, false) \
MSCCL_DECL_KERNEL_ENTRY_FUNC_DEVREDOP(Min, false) \
MSCCL_DECL_KERNEL_ENTRY_FUNC_DEVREDOP(Max, false) \
MSCCL_DECL_KERNEL_ENTRY_FUNC_DEVREDOP(Min, false) \
MSCCL_DECL_KERNEL_ENTRY_FUNC_DEVREDOP(PreMulSum, false) \
MSCCL_DECL_KERNEL_ENTRY_FUNC_DEVREDOP_NOFLOAT(SumPostDiv, false)