SWDEV-422207 - Report kernel names for activity profiling

- Report kernel names for optimized graph path
- Refactor code so that we store profiling info in Accumulate command

Change-Id: Ib97735a0239aeb9fc3a50a4bb7126dd0bcadc8af


[ROCm/clr commit: b056686607]

projects/clr/rocclr/device/rocm/rocvirtual.cpp

@@ -135,7 +135,7 @@ void Timestamp::checkGpuTime() {
       // Avoid profiling data for the sync barrier, in tiny performance tests the first call
       // to ROCr is very slow and that also affects the overall performance of the callback thread
       if (command().GetBatchHead() == nullptr || command().profilingInfo().marker_ts_
-          || command().profilingInfo().multiple_ts_) {
+          || command().type() == CL_COMMAND_TASK) {
         hsa_amd_profiling_dispatch_time_t time = {};
         if (it->engine_ == HwQueueEngine::Compute) {
           hsa_amd_profiling_get_dispatch_time(gpu()->gpu_device(), it->signal_, &time);
@@ -149,8 +149,8 @@ void Timestamp::checkGpuTime() {
         start = std::min(time.start, start);
         end = std::max(time.end, end);
 
-        if (command().profilingInfo().multiple_ts_) {
-          command().AddTimeStamps(time.start, time.end);
+        if (command().type() == CL_COMMAND_TASK) {
+          static_cast<amd::AccumulateCommand&>(command()).addTimestamps(time.start, time.end);
         }
 
         ClPrint(amd::LOG_INFO, amd::LOG_SIG, "Signal = (0x%lx), start = %ld, "