On Sun, 05 Dec 2010 10:13:05 +0200 Tom Hacohen <tom.haco...@partner.samsung.com> said:
i have no problems with the README suggestion optimizations you'd like to use. making it a configure --enable-optimizations isn't any better than saying 'recommended optimizations: export CFLAGS="..."' with the first being more obscure. as such we just should encourage people packaging EFL to use them, but they have their own policies too. we could recommend optimizing for core2 and better but their policy may be to support pentium and up or 486 and up.. or 386 and up - or in the arm world we could say "optimize for armv7!" but they have to support armv4... the linker optimizations are not arch dependent ones of course, but they are toolchain dependent - does your toolchain support it... is it stable? brand new feature of gcc. definitely hasn't been tested widely by now. do we go turning an optimization on just because the compiler CAN do it? inform the packagers etc. that such an optimization may exist and that it may help - sure. do we even know how much it helps? do we have numbers? no. in theory it can help by inlining across objects. great. avoid a call and return for such inlined functions. BUT... instead you increase your instruction fetch bandwidth as you inlined instead of recycled, and your binary sizes go up as you "copy and paste" code around. there comes a price for the optimization. i say that we leave it up to the person compiling to make that choice as even WE don't know for sure if it even helps, and if so, how much, and what the general price in size is that we pay. > I generally agree with this claim, and it's true about many things, but > I also think we should include "recommended defaults" for people who are > unsure about what's good/wanted. If not in the configure so at least in > the README/INSTALL :) > > Don't you think? This will mean more users will benefit from a faster > system (and that's what we are here for...). > > -- > Tom. > > On Sat, 2010-12-04 at 14:15 +0900, Carsten Haitzler wrote: > > On Sun, 28 Nov 2010 12:37:53 +0100 (CET) Vincent Torri <vto...@univ-evry.fr> > > said: > > > > belongs in CFLAGS/LDFLAGS of the people doing the compiling. > > (users/packagers) > > > > > > > > Hey, > > > > > > with newer gcc, we could optimize link with -flto (and > > > also maybe -fwhole-program). Shouldn't we add those options if they are > > > available (i have already written m4 macro to add such flags) ? > > > > > > Vincent > > > > > > ------------------------------------------------------------------------------ > > > Increase Visibility of Your 3D Game App & Earn a Chance To Win $500! > > > Tap into the largest installed PC base & get more eyes on your game by > > > optimizing for Intel(R) Graphics Technology. Get started today with the > > > Intel(R) Software Partner Program. Five $500 cash prizes are up for grabs. > > > http://p.sf.net/sfu/intelisp-dev2dev > > > _______________________________________________ > > > enlightenment-devel mailing list > > > enlightenment-devel@lists.sourceforge.net > > > https://lists.sourceforge.net/lists/listinfo/enlightenment-devel > > > > > > > > > -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- The Rasterman (Carsten Haitzler) ras...@rasterman.com ------------------------------------------------------------------------------ What happens now with your Lotus Notes apps - do you make another costly upgrade, or settle for being marooned without product support? Time to move off Lotus Notes and onto the cloud with Force.com, apps are easier to build, use, and manage than apps on traditional platforms. Sign up for the Lotus Notes Migration Kit to learn more. http://p.sf.net/sfu/salesforce-d2d _______________________________________________ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
