On Tue, Dec 10, 2019 11:13:47 +0100, Florian Weimer wrote: > * Ankur Sinha: > > > -D_GLIBCXX_ASSERTIONS was added to the default build flags in F28[1]. > > For one of my packages, upstream reports that this results in a 25% slow > > down[2], which is quite a lot. > > > > Has anyone else noticed such performance deterioration as a result of this? > > We have seen cases where GCC cannot eliminate the bounds check in a > tight loop because the array index has the wrong type. This is arguably > a bug in the source code; fixing the type also addresses the performance > regression. > > Apart from that, I don't know of any performance regressions. We would > have to look at profiles and disassembly to see what's going on in the > case of arbor-sim. Thanks for your reply, Florian. Upstream have dug in to this. Could you please have a look and suggest what to do in this case? https://github.com/arbor-sim/arbor/issues/917#issuecomment-564106931 -- Thanks, Regards, Ankur Sinha "FranciscoD" (He / Him / His) | https://fedoraproject.org/wiki/User:Ankursinha Time zone: Europe/London
Attachment:
signature.asc
Description: PGP signature
_______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@xxxxxxxxxxxxxxxxxxxxxxx
