rggammon
/
efl
forked from enlightenment/efl
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

evas common: NEON version of evas_common_convert_argb_premul. 

This patch reduces power consumption by around 18mA in certain scenarios
(music player list scroll, my files sound list scroll), making
evas_common_convert_argb_premul() ~60% faster (6.2msec->2.6msec).

Take music-player application, make 100 copies of the standard
Over the Horizon” song, scroll up and down to see those
downscaled-from-720x720 thumbnails enter and leave the screen.
Every time a list item enters the screen, the image is re-read
(as evas image cache is not large enough to store more than two
pictures of that size), and one call of _common_convert_argb_premul()
occurs, taking ~6.2msec (which is not much compared to ~60msec
spent in libpng->libz (the biggest bottleneck here),
but still noticeable).

A similar power consumption improvement is observed during
scrolling sounds list of the same files in My Files application
(just with idle level ~100mA lower).

We also checked the new code to be correct on random input data.

all tests are performed based on tizen device.

Signed-Off-By: Artem Dergachev <dergachev.a@samsung.com>
This commit is contained in:
ChunEon Park 2015-04-07 23:02:15 +09:00
parent 2bbdc17bea
commit 51d60e649c
1 changed files with 40 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
src/lib/evas/common/evas_convert_color.c
View File

@ -1,6 +1,10 @@
#include "evas_common_private.h"
#include "evas_convert_color.h"
#ifdef BUILD_NEON
#include <arm_neon.h>
#endif
EAPI DATA32
evas_common_convert_ag_premul(DATA16 *data, unsigned int len)
{
@ -28,6 +32,42 @@ evas_common_convert_argb_premul(DATA32 *data, unsigned int len)
DATA32 *de = data + len;
DATA32 nas = 0;
#ifdef BUILD_NEON
if (evas_common_cpu_has_feature(CPU_FEATURE_NEON))
{
uint8x8_t mask_0x00 = vdup_n_u8(0);
uint8x8_t mask_0x01 = vdup_n_u8(1);
uint8x8_t mask_0xff = vdup_n_u8(255);
uint8x8_t cmp;
while (data <= de - 8)
{
uint8x8x4_t rgba = vld4_u8(data);
cmp = vand_u8(vorr_u8(
vceq_u8(rgba.val[3], mask_0xff),
vceq_u8(rgba.val[3], mask_0x00)
), mask_0x01);
nas += vpaddl_u32(vpaddl_u16(vpaddl_u8(cmp)));
uint16x8x4_t lrgba;
lrgba.val[0] = vmovl_u8(rgba.val[0]);
lrgba.val[1] = vmovl_u8(rgba.val[1]);
lrgba.val[2] = vmovl_u8(rgba.val[2]);
rgba.val[0] = vshrn_n_u16(vmlal_u8(lrgba.val[0], rgba.val[0], rgba.val[3]), 8);
rgba.val[1] = vshrn_n_u16(vmlal_u8(lrgba.val[1], rgba.val[1], rgba.val[3]), 8);
rgba.val[2] = vshrn_n_u16(vmlal_u8(lrgba.val[2], rgba.val[2], rgba.val[3]), 8);
vst4_u8(data, rgba);
data += 8;
}
}
#endif
while (data < de)
{
DATA32 a = 1 + (*data >> 24);