On 11/23/2016 07:33 PM, Zhengyu Gu wrote: > Webrev: > http://cr.openjdk.java.net/~zgu/shenandoah/par-om/webrev.00/index.html Ok, good! Can you test this with lower number of monitors, e.g. set -p count=1,10,100,1000? This will assert we don't have the regression for the usual case. Thanks, -Aleksey