Use barrier packets for event profiling

Use barrier packets for every profile marker that gets submitted
and use the completion signal to get GPU ts. This gives most accurate
dispatch time. Club cache flushes with profile marker if there is a
pending dispatch that needs cache flush. This optimization saves on
extra barrier and helps wall time

Change-Id: Ib62d6d7aabf4743827b561be6c9c5afa813203da


[ROCm/clr commit: 59c6cb0268]

projects/clr/rocclr/platform/command.hpp

@@ -97,7 +97,7 @@ class Event : public RuntimeObject {
   static const EventWaitList nullWaitList;
 
   struct ProfilingInfo {
-    ProfilingInfo(bool enabled = false) : enabled_(enabled), waves_(0) {
+    ProfilingInfo(bool enabled = false) : enabled_(enabled), waves_(0), marker_ts_(false) {
       if (enabled) {
         clear();
         callback_ = nullptr;
@@ -111,6 +111,7 @@ class Event : public RuntimeObject {
     bool enabled_;    //!< Profiling enabled for the wave limiter
     uint32_t waves_;  //!< The number of waves used in a dispatch
     ProfilingCallback* callback_;
+    bool marker_ts_;
     void clear() {
       queued_ = 0ULL;
       submitted_ = 0ULL;