
4 Jul
2009
4 Jul
'09
8:52 a.m.
I've been investigating why float_sort is so much faster than std::sort on x86 processors, and why sorting integers is so much faster than float_sort, and came to the conclusion that floating-point comparisons are extremely slow on x86.
Hey Steven, Is it true even with SSE3 optimizations turned on and data aligned to 16-bytes boundaries? Which compiler were you using? Regards. -- Edouard __________ Information from ESET NOD32 Antivirus, version of virus signature database 4215 (20090704) __________ The message was checked by ESET NOD32 Antivirus. http://www.eset.com