[x264-devel] [PATCH] Add all remaining 16x16 predict Altivec routines

[Guillaume POIRIER]

Hello folks,

Attached patch add Altivec implementation of all the remaining 16x16
routines.

I haven't yet benched it as don't have a machine on which I can enable
the hardware counters to perform some precise time measurement (my PPC
is still at the repair store).

I believe it's as optimal as it can get with Altivec (modulo some
unrolling off course), so either a given routine is faster than the C
version and it should be enabled, or it's slower and it shouldn't.

Thoughts? Benchmark figures?

Guillaume
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: predict_altivec_16x16.diff
Url: http://mailman.videolan.org/pipermail/x264-devel/attachments/20090112/7e7bd912/attachment.txt