Gerrit Voss wrote: > Hi, > > On Sun, 2007-09-30 at 11:08 +0200, Andreas Zieringer wrote: > >> couldn't reproduce this I get the following output with your test program: >> >> CPU cycles with manual inlining: 758135486 >> CPU cycles With Pointinterface: 811406783 >> Ratio: 1.07027 > > Just a quick update, I ran some more test, same compiler options, > VS2005: > > original opensg: > > Ratio: 1.32355 > > > hand coded Vec3f class (with __forceinline): > > Ratio: 1.29539 > > > gmtl (even more templates) : > > Ratio: 0.992 - 1.0003 > > > > So more templates are better, I knew it ;-). Ok, I guess we have to do > some more tests and have a closer look ;-)
One thing to point out in the interest of fairness is that we have not done any detailed tuning in GMTL (yet). The code "should" perform fairly well, but we have not gone through and made sure that everything is near optimal and we definitely haven't done anything to the point of the blitz library to make sure that expression templates create code that compilers can optimize (and in some cases even creates SIMD code on supported hardware). The nice thing is that this type of support is possible and could "automagically" make user code much faster. -Allen > > > > kind regards, > gerrit > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Microsoft > Defy all challenges. Microsoft(R) Visual Studio 2005. > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ > _______________________________________________ > Opensg-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/opensg-users > ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2005. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Opensg-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/opensg-users
