Merge "Add neon optimize iht8x8 which is 282% faster than C."

86fb12b6 · Hangyu Kuang · Gerrit Code Review · 25655e57 · 182366c7 · 86fb12b6
Commit 86fb12b6 authored 11 years ago by Hangyu Kuang Committed by Gerrit Code Review 11 years ago
--- a/vp9/common/arm/neon/vp9_short_iht8x8_add_neon.asm
+++ b/vp9/common/arm/neon/vp9_short_iht8x8_add_neon.asm
--- a/vp9/common/vp9_rtcd_defs.sh
+++ b/vp9/common/vp9_rtcd_defs.sh
@@ -328,7 +328,7 @@ prototype void vp9_short_iht4x4_add "int16_t *input, uint8_t *dest, int dest_str
 specialize vp9_short_iht4x4_add sse2 neon

 prototype void vp9_short_iht8x8_add "int16_t *input, uint8_t *dest, int dest_stride, int tx_type"
-specialize vp9_short_iht8x8_add sse2
+specialize vp9_short_iht8x8_add sse2 neon

 prototype void vp9_short_iht16x16_add "int16_t *input, uint8_t *output, int pitch, int tx_type"
 specialize vp9_short_iht16x16_add sse2

--- a/vp9/vp9_common.mk
+++ b/vp9/vp9_common.mk
@@ -106,6 +106,7 @@ VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_short_idct16x16_1_add_neon$(
 VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_short_idct16x16_add_neon$(ASM)
 VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_short_idct32x32_add_neon$(ASM)
 VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_short_iht4x4_add_neon$(ASM)
+VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_short_iht8x8_add_neon$(ASM)
 VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_mb_lpf_neon$(ASM)
 VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_copy_neon$(ASM)
 VP9_COMMON_SRCS-$(HAVE_NEON) += common/arm/neon/vp9_avg_neon$(ASM)