[Bug 3323] math routine performance tweaks

bugzilla-daemon at icculus.org bugzilla-daemon at icculus.org
Thu Aug 23 17:20:28 EDT 2007


http://bugzilla.icculus.org/show_bug.cgi?id=3323





------- Comment #6 from devhc97 at gmail.com  2007-08-23 17:20 EDT -------
Created an attachment (id=1486)
 --> (http://bugzilla.icculus.org/attachment.cgi?id=1486&action=view)
RotatePointAroundVector() accuracy test and performance test

This is the program I used to test the function, now I've included all three
versions in it for comparison.

On my compiler, using float, and fast fp model, the functions do differ in
accuracy a bit: they're accurate to ~6 digits, so basically I think they differ
only as much as the float's epsilon. Performance run: Q3=23373ms, TREM=11186ms,
NEW=10235ms(10% faster).

OK, so if this will cause prediction misses with higher values, then maybe set
this to REMIND, and add it in an ioq3 point release.


-- 
Configure bugmail: http://bugzilla.icculus.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug, or are watching the QA contact.



More information about the quake3-bugzilla mailing list