2.16.2-1

Add support for CUDA 12.0, drop Kepler (sm_35).
Support for H100 features.
Make socket code more robust and protected. Solves #555.
Improve performance on large CUDA graphs, reducing dependencies.
Reduce inter-socket bandwidth on AMD CPUs to favor better paths.
Various fixes to ncclCommAbort.
Make service thread polling resistant to EINTR.
Compile with profiling API by default.
Extend NVTX instrumentation with call arguments.

src/include/shm.h

@@ -9,7 +9,9 @@
 
 #include "nccl.h"
 
-ncclResult_t ncclShmOpen(char* shmPath, const int shmSize, void** shmPtr, void** devShmPtr, int create);
-ncclResult_t ncclShmUnlink(const char* shmname);
-ncclResult_t ncclShmClose(void* shmPtr, void* devShmPtr, const int shmSize);
+typedef void* ncclShmHandle_t;
+ncclResult_t ncclShmOpen(char* shmPath, size_t shmSize, void** shmPtr, void** devShmPtr, int refcount, ncclShmHandle_t* handle);
+ncclResult_t ncclShmClose(ncclShmHandle_t handle);
+ncclResult_t ncclShmUnlink(ncclShmHandle_t handle);
+
 #endif