sbrdsp: Unroll and use integer operations
authorChristophe Gisquet <christophe.gisquet@gmail.com>
Tue, 9 Apr 2013 21:57:10 +0000 (21:57 +0000)
committerDiego Biurrun <diego@biurrun.de>
Fri, 3 May 2013 16:23:30 +0000 (18:23 +0200)
commit4a7af92cc80ced8498626401ed21f25ffe6740c8
treeee346d6cd816512e4527e13edbd0e158cfa13d71
parent8394d9a676db1ffa1c8e4f6dd239e154d051ce03
sbrdsp: Unroll and use integer operations

This patch can be controversial, by assuming floats are IEEE-754 and
particular behaviour of the FPU will get in the way.
Timing on Arrandale and Win32 (thus, x87 FPU is used in the reference).

sbr_qmf_pre_shuffle_c: 115 to 76
sbr_neg_odd_64_c: 84 to 55
sbr_qmf_post_shuffle_c: 112 to 83

Signed-off-by: Diego Biurrun <diego@biurrun.de>
libavcodec/sbrdsp.c