2.11.4-1

Add new API for creating a reduction operation which multiplies the input by a rank-specific scalar before doing an inter-rank summation (see: ncclRedOpCreatePreMulSum). Improve CollNet (SHARP) performance of ncclAllReduce when captured in a CUDA Graph via user buffer registration. Add environment variable NCCL_NET_PLUGIN="<suffix>" to allow user to choose among multiple NCCL net plugins by substituting into "libnccl-net-<suffix>.so". Fix memory leak of NVB connections. Fix topology detection of IB Virtual Functions (SR-IOV). [ROCm/rccl commit: e11238b302]
2021-09-08 13:56:25 -07:00
parent 04553b802a
commit e51f2a83da
41 changed files with 1532 additions and 599 deletions
@@ -10,6 +10,7 @@
 #include "nccl.h"
 #include <stdint.h> // for standard [u]intX_t types
 #include <stdio.h>
+#include <stdlib.h>

 // These can be used if the GDR library isn't thread safe
 #include <pthread.h>