Merge remote-tracking branch 'remotes/nccl/master' into HEAD

[ROCm/rccl commit: 7c38da0939]

projects/rccl/pkg/redhat/nccl.spec.in

@@ -7,6 +7,7 @@ Group:          Development/Libraries
 License:        BSD
 URL:            http://developer.nvidia.com/nccl
 Source0:        nccl_${nccl:Major}.${nccl:Minor}.${nccl:Patch}${nccl:Suffix}-${pkg:Revision}+cuda${cuda:Major}.${cuda:Minor}_${pkg:Arch}.txz
+Prereq:         /sbin/ldconfig
 
 %description
 NCCL (pronounced "Nickel") is a stand-alone library of standard collective
@@ -50,6 +51,12 @@ ln -s libnccl.so.${nccl:Major} $RPM_BUILD_ROOT/%{_libdir}/libnccl.so
 # static
 install -m 644 lib/libnccl_static.a $RPM_BUILD_ROOT/%{_libdir}
 
+%post -p /sbin/ldconfig
+%postun -p /sbin/ldconfig
+
+%post devel -p /sbin/ldconfig
+%postun devel -p /sbin/ldconfig
+
 %clean
 rm -rf $RPM_BUILD_ROOT
 

projects/rccl/src/include/socket.h

@@ -327,7 +327,11 @@ static ncclResult_t createListenSocket(int *fd, union socketAddress *localAddr)
   if (socketToPort(&localAddr->sa)) {
     // Port is forced by env. Make sure we get the port.
     int opt = 1;
+#if defined(SO_REUSEPORT)
     SYSCHECK(setsockopt(sockfd, SOL_SOCKET, SO_REUSEADDR | SO_REUSEPORT, &opt, sizeof(opt)), "setsockopt");
+#else
+    SYSCHECK(setsockopt(sockfd, SOL_SOCKET, SO_REUSEADDR, &opt, sizeof(opt)), "setsockopt");
+#endif
   }
 
   // localAddr port should be 0 (Any port)

projects/rccl/src/include/utils.h

@@ -12,7 +12,7 @@
 #include <stdint.h>
 
 ncclResult_t getHostName(char* hostname, int maxlen, const char delim);
-uint64_t getnHash(const char* string, int n);
+uint64_t getHash(const char* string, int n);
 uint64_t getHostHash();
 uint64_t getPidHash();
 

projects/rccl/src/init.cc

@@ -765,8 +765,8 @@ static ncclResult_t initTransportsRank(struct ncclComm* comm, ncclUniqueId* comm
 
   int rank = comm->rank;
   int nranks = comm->nRanks;
-  uint64_t commHash = getnHash(commId->internal, NCCL_UNIQUE_ID_BYTES);
-  TRACE(NCCL_INIT, "comm %p, commHash %lu, rank %d nranks %d - BEGIN", comm, commHash, rank, nranks);
+  uint64_t commHash = getHash(commId->internal, NCCL_UNIQUE_ID_BYTES);
+  TRACE(NCCL_INIT, "comm %p, commHash %lx, rank %d nranks %d - BEGIN", comm, commHash, rank, nranks);
   NCCLCHECK(bootstrapInit(commId, rank, nranks, &comm->bootstrap));
 
   // AllGather1 - begin

projects/rccl/src/misc/utils.cc

@@ -88,10 +88,10 @@ void ncclDebugLog(ncclDebugLogLevel level, unsigned long flags, const char *file
   }
 }
 
-uint64_t getHash(const char* string) {
+uint64_t getHash(const char* string, int n) {
   // Based on DJB2, result = result * 33 + char
   uint64_t result = 5381;
-  for (int c = 0; string[c] != '\0'; c++) {
+  for (int c = 0; c < n; c++) {
     result = ((result << 5) + result) + string[c];
   }
   return result;
@@ -130,7 +130,7 @@ uint64_t getHostHash(void) {
   uname[offset]='\0';
   TRACE(NCCL_INIT,"unique hostname '%s'", uname);
 
-  return getHash(uname);
+  return getHash(uname, strlen(uname));
 }
 
 /* Generate a hash of the unique identifying string for this process
@@ -150,7 +150,7 @@ uint64_t getPidHash(void) {
   pname[plen+len]='\0';
   TRACE(NCCL_INIT,"unique PID '%s'", pname);
 
-  return getHash(pname);
+  return getHash(pname, strlen(pname));
 }
 
 int parseStringList(const char* string, struct netIf* ifList, int maxList) {