On Mon, 23 Sep 2013, Andreas Hansson wrote:
On Sept. 21, 2013, 3:39 p.m., Andreas Hansson wrote:
src/sim/eventq_impl.hh, line 55
<http://reviews.gem5.org/r/1667/diff/9/?file=37731#file37731line55>
is it worth checking inParallelMode first?
As a side not, do we know what the performance impact is when not running
in parallel mode? (perhaps this has been resolved already)
Nilay Vaish wrote:
I'll run the regression tests and report the execution time differences
soon.
Andreas Hansson wrote:
My _very_ early results suggest around -15% host_inst_rate for most of the
regressions, but I'll get back to you tomorrow with a more conclusive result.
Andreas Hansson wrote:
Ok, so a full regression run, including compilation of all ISAs is now 18%
slower. This is a fairly accurate measure of the slowdown, as the run time is
within +-10 min over the last months.
This suggests actual simulation is more than 18% slower, which obviously is
a _big_ issue. I haven't done any profiling, but will keep you posted on any
findings.
Steve Reinhardt wrote:
That is painful. I hope that we can identify the source of the overhead
and solve it with (at worst) some special-casing on (numEventQueues == 1) that
isn't too intrusive or ugly.
Thanks for thinking of this... it didn't even occur to me to check.
I've got a second point now as well, and after fixing the PC regressions that
were failing (see remark below), the total runtime of the regressions is now up
19% (including a clean compile of all ISAs). I haven't done any profiling as of
yet.
So I just ran the regression test: 30.eon for alpha architecture, five
times with and with out the patch. I am seeing improved times for the
simulation when the patch is applied.
Original: 3m 23.54s, 3m 23.405s, 3m 23.235s, 3m 23.925s, 3m 23.138s
With the patch: 3m 3.39s, 2m 59.65s, 3m 1.01s, 3m 1.499s, 3m 1.004s
I read the patch and I am not able to come up with reasons why the times
should change considerably (unless tls has a significant affect).
--
Nilay
_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev
