This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [Patch] Step2: don't use _M_out_lim in filebuf, have filebufnon unified, etc...


Paolo Carlini wrote:

Here, if both are moved 9876 regresses to the pre-patch performance...

... however, adding -finline-limit=300 to the command line (that is _reducing_ it from its default 300), restores the performance... It looks like, for some reason, in a tight loop with only sputc, there is an over-inlining problem...


Anyway, those weirdness should be fixed with Honza work and, basing on general principles, I think we should move in any case both sputc and sbumpc inline. Do you agree? What about sungetc and sputbackc?

Paolo.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]