pmc: add parraller kernels 'rocprof' option

The profiling was only enabled in serial mode, i.e., kernels
are serialized in execution, and counters are reset at each
kernel start and read at kernel completion. This patch adds
the concurrent mode, by issuing the process-level start
packet to reset counters, and then reading twice at kernel
start and end time to obtain the counter value difference.
The new concurrent profiling usage needs the integration
with the corresponding augment at aqlprofile side.

Change-Id: I94b4442eadc8c64b8fba51b1e4916fc8b895ad21

enable cnocurrent kernels 'rocprof' option

Change-Id: I5088d54879a4bf91e180eba520aa9427304c1713

bin/rpl_run.sh

@@ -183,6 +183,7 @@ usage() {
   echo "    Supported time formats: <number(m|s|ms|us)>"
   echo "  --flush-rate <rate> - to enable trace flush rate (time period)"
   echo "    Supported time formats: <number(m|s|ms|us)>"
+  echo "  --parallel-kernels - to enable cnocurrent kernels"
   echo ""
   echo "Configuration file:"
   echo "  You can set your parameters defaults preferences in the configuration file 'rpl_rc.xml'. The search path sequence: .:${HOME}:<package path>"
@@ -439,6 +440,9 @@ while [ 1 ] ; do
     if [ "$2" = "off" ] ; then
       export ROCP_OBJ_TRACKING=0
     fi
+  elif [ "$1" = "--parallel-kernels" ] ; then
+    ARG_VAL=0
+    export ROCP_K_CONCURRENT=1
   elif [ "$1" = "--verbose" ] ; then
     ARG_VAL=0
     export ROCP_VERBOSE_MODE=1