I wonder if it was that massive looking commit where you separated out the coherence logic. I am currently running STREAM on an old commit (the time stats remodel commit) and it appears to be working fine. I'll try to do another git bisect with STREAM this time since it seems to be a better determinant of failure.
On Tue, Jul 12, 2011 at 12:39 PM, avadh patel <[email protected]> wrote: > I am also having some issues with multicore setup of Atom. I am running > parsec benchmarks and only couple of them run successfully. I am trying to > debug the issue and have no idea what's wrong with the Atom core model. All > the single-core benchmarks runs just fine but multicore is having trouble. > If you can share your test kernel then I can try to debug with it. > > - Avadh > > On Tue, Jul 12, 2011 at 9:09 AM, DRAM Ninjas <[email protected]> wrote: > >> ok after running a different workload it appears that I do, in fact, have >> the wrong commit, but I am still having a problem with later commits. >> >> >> On Tue, Jul 12, 2011 at 11:48 AM, DRAM Ninjas <[email protected]>wrote: >> >>> Hello everyone, >>> >>> I've updated my core-models branch recently and I have been getting >>> assertion failures that there have been no commits. The workload is a super >>> simple little C test program I wrote and it errors out with the assertion >>> failure almost immediately. I have been using -machine shared_l2 with 2 atom >>> cores. I was wondering if anyone else has had any issues? >>> >>> I did a git bisect and it appears the first bad commit is: >>> >>> commit 35b044d6a47a9decc07ee9cfecac63ac72148902 >>> Author: Avadh Patel <[email protected]> >>> Date: Mon Jul 4 12:15:13 2011 -0700 >>> >>> Small bug fix in mesi logic >>> >>> Except this commit is trivial -- so either I got lucky on certain commits >>> that I marked good (my workload is rather short) and ended up getting the >>> wrong culprit, or this change has bigger implications somewhere else. I just >>> wanted to see if anyone else has had issues before I start trying to debug >>> this ... >>> >>> -Paul >>> >> >> >> _______________________________________________ >> http://www.marss86.org >> Marss86-Devel mailing list >> [email protected] >> https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel >> >> >
_______________________________________________ http://www.marss86.org Marss86-Devel mailing list [email protected] https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
