On Tue, 2008-12-09 at 14:29 +0530, Sayamindu Dasgupta wrote:
>
> There's a cairo configure option:
>
> --disable-some-floating-point
> Disable certain code paths that rely heavily on
> double precision floating-point calculation. This
> option can improve performance on systems without a
> double precision floating-point unit, but might
> degrade performance on those that do.
>
>
> Maybe someone can do some tests with this flag on. I'm totally
> clueless about the capabilities of the XO CPU, but it might be an
> interesting thing to try out.
May be worth trying, but the last I knew it didn't do anything (see
attached), but was present primarily to aid people who wanted to
experiment with such code paths.
- Jim
--
Jim Gettys <[EMAIL PROTECTED]>
One Laptop Per Child
--- Begin Message ---
On Fri, 14 Mar 2008 12:04:02 -0700, Marc Singer wrote:
> I'll look again. The 1.2.4 had a 'remove-some-floating-point' option.
> There may be some new ones in the later releases.
Right. We added a "--disable-some-floating-point" option, (or similar
name), when we found some code that improved performance on
non-hardware-floating-point systems but regressed performance on
systems with hardware floating-point.
*But* shortly after adding the option, (and before the first release
that included it), we found an alternate way to write the affected
code that was faster for both kinds of systems. So, so far, the option
doesn't actually change anything, (as far as I'm aware).
But if you are using a system with very poor performance for
floating-point operations then it certainly doesn't hurt to add that
option to your build configuration. And perhaps in the future it will
change something significant.
But for now, I don't think you'll find it affecting your bug.
Have fun with cairo!
-Carl
pgp03oVVDschK.pgp
Description: PGP signature
_______________________________________________
cairo mailing list
[EMAIL PROTECTED]
http://lists.cairographics.org/mailman/listinfo/cairo
--- End Message ---
_______________________________________________
Sugar-devel mailing list
[email protected]
http://lists.sugarlabs.org/listinfo/sugar-devel
