[Launch] Enable Implicit order launch with serial mode (#2033)

[ROCm/rccl commit: a9bb7e9807]

projects/rccl/src/enqueue.cc

@@ -1586,7 +1586,7 @@ static ncclResult_t getImplicitOrder(enum ncclImplicitOrder *mode, bool capturin
     if (capturing && driver < 12090) { *mode = ncclImplicitOrderSerial; return ncclSuccess; }
     *mode = 12030 <= std::min<int>(CUDART_VERSION, driver) ? ncclImplicitOrderLaunch : ncclImplicitOrderSerial;
 #else
-    *mode = ncclImplicitOrderNone;
+    *mode = ncclImplicitOrderSerial;
 #endif
     return ncclSuccess;
   }
@@ -1900,10 +1900,10 @@ ncclResult_t ncclLaunchFinish(struct ncclComm* comm) {
     ncclIntruQueueConstruct(&planner->planQueue);
 
     bool capturing = ncclCudaGraphValid(planner->capturingGraph);
-    //cudaStream_t launchStream = planner->streams->stream; // First user stream gets launch // unused variable - compiler warning
+    cudaStream_t launchStream = planner->streams->stream; // First user stream gets launch
     cudaStream_t deviceStream, launchOrder;
-
     cudaEvent_t finishedEvent = comm->sharedRes->scratchEvent;
+    CUDACHECK(cudaEventRecord(finishedEvent, launchStream));
 
     if (comm->workFifoProduced - comm->workFifoProducedLastRecorded > comm->workFifoBytes/8) {
       comm->workFifoProducedLastRecorded = comm->workFifoProduced;
@@ -1918,8 +1918,6 @@ ncclResult_t ncclLaunchFinish(struct ncclComm* comm) {
     }
 
     if (capturing || planner->numStreams != 1) {
-      // CUDACHECK(cudaEventRecord(finishedEvent, launchStream));
-
       // deviceStream waits on userStream[0]
       NCCLCHECK(ncclStrongStreamAcquiredWorkStream(planner->capturingGraph, &comm->sharedRes->deviceStream, /*concurrent=*/false, &deviceStream));
 

projects/rccl/src/include/rocmwrap.h

@@ -9,6 +9,7 @@
 #define NCCL_ROCMWRAP_H_
 
 #include <hsa/hsa.h>
+#include "checks.h"
 
 typedef hsa_status_t (*PFN_hsa_init)();
 typedef hsa_status_t (*PFN_hsa_system_get_info)(hsa_system_info_t attribute, void* value);
@@ -85,6 +86,17 @@ extern CUmemAllocationHandleType ncclCuMemHandleType;
 
 ncclResult_t rocmLibraryInit(void);
 
+extern int ncclCudaDriverVersionCache;
 extern bool ncclCudaLaunchBlocking; // initialized by ncclCudaLibraryInit()
 
+inline ncclResult_t ncclCudaDriverVersion(int* driver) {
+  int version = __atomic_load_n(&ncclCudaDriverVersionCache, __ATOMIC_RELAXED);
+  if (version == -1) {
+    CUDACHECK(cudaDriverGetVersion(&version));
+    __atomic_store_n(&ncclCudaDriverVersionCache, version, __ATOMIC_RELAXED);
+  }
+  *driver = version;
+  return ncclSuccess;
+}
+
 #endif

projects/rccl/src/misc/rocmwrap.cc

@@ -28,6 +28,8 @@ DECLARE_ROCM_PFN(hsa_status_string);
 
 static void *hsaLib;
 static uint16_t version_major, version_minor;
+
+int ncclCudaDriverVersionCache = -1;
 bool ncclCudaLaunchBlocking = false;
 
 static pthread_once_t initOnceControl = PTHREAD_ONCE_INIT;