Hi all,

Recently I simulated 4-core alpha running Parsec benchmark in FS mode. And
in the output stats.txt, I observed a large number of quiesceCycles for
each core in ROI. I learned from the previous thread that the quiesceCycles
are caused by interrupts. But I want to link the quiesceCycles with the
benchmark features and understand the behavior. For example, from the
Parsec documentation I found there is a large number of
locks/barriers-synchronizations in some Parsec programs. Is this the reason
that causes the large number of quiesceCycles?

Thanks a lot in advance!
gem5-users mailing list

Reply via email to