ARM: NEON: Instruction scheduling of bilinear over_8888_8888
authorTaekyun Kim <tkq.kim@samsung.com>
Wed, 21 Sep 2011 06:52:13 +0000 (15:52 +0900)
committerTaekyun Kim <tkq.kim@samsung.com>
Tue, 18 Oct 2011 04:14:28 +0000 (13:14 +0900)
commit4481920f405e47b3a92811a8cb06afbd37dee01b
treeedb867e0661cb5d9c78ee273529351485e1ee070
parent1cd916f3a5ebeb943f66eecf0b8ce99af0b95d11
ARM: NEON: Instruction scheduling of bilinear over_8888_8888

Instructions are reordered to eliminate pipeline stalls and get
better memory access.

Performance of before/after on cortex-a8 @ 1GHz

<< 2000 x 2000 with scale factor close to 1.x >>
before : 50.43 Mpix/s
after  : 61.09 Mpix/s
pixman/pixman-arm-neon-asm-bilinear.S