SWDEV-538195 - Introduce threshold for handler submission (#723)

- When doing device/stream sync, we can submit a handler which may
  introduce some host side delays. Use DEBUG_CLR_BATCH_CPU_SYNC_SIZE to
  batch commands for host wait. Default for HIP is 8 commands.
- Investigation is underway in ROCr but need to address this for now in
  HIP runtime.

rocclr/platform/commandqueue.hpp

@@ -251,6 +251,9 @@ class HostQueue : public CommandQueue {
   //! Get the submitted batch
   Command* GetSubmissionBatch() const { return head_; }
 
+  //! Get the current batch size
+  size_t GetSubmissionBatchSize() const { return size_; }
+
   //! Insert a command into the linked list of submitted commands
   void FormSubmissionBatch(Command* command) {
     // Insert the command to the linked list.