Closed simoatze closed 7 years ago
The changes look good to me. Travis for the initial PR failed, so I rerun it and it succeeded. I actually don't understand, why it failed in the specific test case.
Simone, please decide on Dong's wording suggestions and merge the pull request then.
@dongahn it does not make me accept the review, I already made those changes though, please merge if everything looks good to you. Thanks!
@simoatze: did you mean to change the wording for flush flag as well?
Our experiments show that this can reduce memory overhead by ~30% and runtime overhead by ~10%. This flag is useful for large OpenMP applications that typically require large amounts of memory, causing out-of-memory exceptions when checked by Archer.
@dongahn Yeah I put your suggestion, is that ok?
Sorry... what is the commit? I actually don't see that change.
I thought I did it, can you see it now?
Hmmm. Maybe I'm looking at the wrong commits. Is 71f4f7 the latest?
Yeah, I see those now. Would it possible to squash intermediate commits? (rebase -i is your tool :-)
@dongahn I rebased, but I think I squashed too much, it should be good though, let me know if it's ok.
@dongahn I made a mess rebasing and now does not let me squash the message, I should probably close this PR and make a new one.
Sure. Thanks. BTW, try to limit the commit title to 50 characters and add the main description into the commit message body.
Standardized flags management to activate debug prints (counters and max memory) and reduce memory overhead in between independent parallel regions.