Fix strange high sample counts in flame graph

Seems to have be because of this bug:
brendangregg/FlameGraph#165. One workaround
there was to record data with `-e cpu-clock`. I don't know what that
means but that is what this commit does.

recordflame.sh

@@ -4,5 +4,5 @@ set -e
 freq=99
 sleep=30
 
-perf record -o perf.data -a -g -F $freq -- sleep $sleep
+perf record -o perf.data -a -g -F $freq -e cpu-clock -- sleep $sleep
 perf script -i perf.data | ./stackcollapse-perf.pl | ./flamegraph.pl --title "$1,freq=$freq,sleep=$sleep" --hash