coffee/libyuv
1
0
Fork 0
mirror of https://chromium.googlesource.com/libyuv/libyuv synced 2025-12-08 01:36:47 +08:00
Code Issues Packages Projects Releases Wiki Activity

[AArch64] Improve RGB565TOARGB using SRI instructions

Browse Source 
The existing code performs a lot of shifts and combines the R and B
components into a single vector unnecessarily. We can express this much
more cleanly by making use of the SRI instruction to insert and replace
shifted bits into the original data, performing the 5/6-bit to 8-bit
expansion in a single instruction if the source bits are already in the
high bits of the byte. We still need a single separate XTN instruction
to narrow the B component before the left shift since Neon does not have
a narrowing left shift instruction.

Reduction in runtime for selected kernels:

               Kernel | Cortex-A55 | Cortex-A76 | Cortex-X2
    RGB565ToYRow_NEON |     -22.1% |     -23.4% |    -25.1%
   RGB565ToUVRow_NEON |     -26.8% |     -20.5% |    -18.8%
 RGB565ToARGBRow_NEON |     -38.9% |     -32.0% |    -23.5%

Bug: libyuv:976
Change-Id: I77b8d58287b70dbb9549451fc15ed3dd0d2a4dda
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/5374286
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Reviewed-by: Justin Green <greenjustin@google.com>
This commit is contained in:
George Steed 2024-03-13 06:36:29 +00:00 committed by Frank Barchard
parent 1ca7c4e1cc
commit 90070986ae
1 changed files with 9 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
source/row_neon64.cc
View File

@ -1611,18 +1611,15 @@ void RAWToRGB24Row_NEON(const uint8_t* src_raw, uint8_t* dst_rgb24, int width) {
);
}
#define RGB565TOARGB \
"shrn v6.8b, v0.8h, #5 \n" /* G xxGGGGGG */ \
"shl v6.8b, v6.8b, #2 \n" /* G GGGGGG00 upper 6 */ \
"ushr v4.8b, v6.8b, #6 \n" /* G 000000GG lower 2 */ \
"orr v1.8b, v4.8b, v6.8b \n" /* G */ \
"xtn v2.8b, v0.8h \n" /* B xxxBBBBB */ \
"ushr v0.8h, v0.8h, #11 \n" /* R 000RRRRR */ \
"xtn2 v2.16b,v0.8h \n" /* R in upper part */ \
"shl v2.16b, v2.16b, #3 \n" /* R,B BBBBB000 upper 5 */ \
"ushr v0.16b, v2.16b, #5 \n" /* R,B 00000BBB lower 3 */ \
"orr v0.16b, v0.16b, v2.16b \n" /* R,B */ \
"dup v2.2D, v0.D[1] \n" /* R */
#define RGB565TOARGB \
/* Input: v0.8h: RRRRRGGGGGGBBBBB */ \
"shrn v1.8b, v0.8h, #3 \n" /* G GGGGGGxx */ \
"shrn v2.8b, v0.8h, #8 \n" /* R RRRRRxxx */ \
"xtn v0.8b, v0.8h \n" /* B xxxBBBBB */ \
"sri v1.8b, v1.8b, #6 \n" /* G GGGGGGGG, fill 2 */ \
"shl v0.8b, v0.8b, #3 \n" /* B BBBBB000 */ \
"sri v2.8b, v2.8b, #5 \n" /* R RRRRRRRR, fill 3 */ \
"sri v0.8b, v0.8b, #5 \n" /* R BBBBBBBB, fill 3 */
void RGB565ToARGBRow_NEON(const uint8_t* src_rgb565,
uint8_t* dst_argb,