[Pixman] [PATCH 0/5] ARM: NEON: Pipelined scaled bilinear functions

[Taekyun Kim]

From: Taekyun Kim <tkq.kim at samsung.com>

Hi, all

Current bilinear NEON codes in pixman-arm-neon-asm-bilinear.S have
still lots of room for improvement. head/tail_head/tail scheme is
applied and I wrote pipelined over_8888_8888 and over_8888_8_8888
bilinear functions. Passed make check and got some speed up. Some
portion of speed up came from adding missed cache prefetch in the
first patch.

Measured on cortex-a8 running @ 1.0GHz

- over_8888_8888 -
before : 35.73 Mpix/s
after  : 61.81 Mpix/s

- over_8888_8_8888 -
before : 30.57 Mpix/s
after  : 51.79 Mpix/s

--
Best Regards,
Taekyun Kim

Taekyun Kim (5):
  ARM: NEON: Some cleanup of bilinear scanline functions
  ARM: NEON: Bilinear macro template for instruction scheduling
  ARM: NEON: Replace old bilinear scanline generator with new template
  ARM: NEON: Instruction scheduling of bilinear over_8888_8888
  ARM: NEON: Instruction scheduling of bilinear over_8888_8_8888

 pixman/pixman-arm-neon-asm-bilinear.S |  922 +++++++++++++++++++++++++++------
 1 files changed, 760 insertions(+), 162 deletions(-)