Hi all. I think that before release we must select what GCC options for user-level software compilation we will set. I played with -march and -mtune, and discovered that Dhrystone test compiled with -march=i486 -mtune=i686 (i486 instruction set, generated code is optimized for execution on i686) is comparable by speed to -march=i686 on k8, and even faster than -march=i686 on k10 (i run it on k10 under x86_64 kernel in x86 emulation that may affect result). I added dry2 package that provides executables compiled for i386, i486, i486 optimized for i686, i586 and i686. Can anybody run them and check if these options will be acceptable for all configurations? Also it'll be good if somebody proposes alternate tests.
------------------------------------------------------------------------------ Nokia and AT&T present the 2010 Calling All Innovators-North America contest Create new apps & games for the Nokia N8 for consumers in U.S. and Canada $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store http://p.sf.net/sfu/nokia-dev2dev _______________________________________________ leaf-devel mailing list leaf-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/leaf-devel
