Merge sendrecv kernel from NCCL 2.7.3

This commit was cherry-picked and modified from
https://github.com/NVIDIA/nccl/commit/5949d96f36d050e59d05872f8bbffd2549318e95

src/include/collectives.h

@@ -82,7 +82,7 @@ DECL_ALL_COLLS
 #define BROADCAST_CHUNKSTEPS 1
 #define REDUCE_SLICESTEPS 1
 #define REDUCE_CHUNKSTEPS 1
-#define SENDRECV_SLICEFACTOR 4
+#define SENDRECV_SLICEFACTOR 1
 #define GATHER_SLICESTEPS 4
 #define GATHER_CHUNKSTEPS 4
 #define SCATTER_SLICESTEPS 4