Commit d203a91a authored by Linfeng Zhang's avatar Linfeng Zhang Committed by Gerrit Code Review

Merge "Add vpx_scaled_2d_neon()"

parents 87846415 9d0d13e9
......@@ -1199,7 +1199,7 @@ const ConvolveFunctions convolve8_neon(
vpx_convolve8_avg_horiz_neon, vpx_convolve8_vert_neon,
vpx_convolve8_avg_vert_neon, vpx_convolve8_neon, vpx_convolve8_avg_neon,
vpx_scaled_horiz_c, vpx_scaled_avg_horiz_c, vpx_scaled_vert_c,
vpx_scaled_avg_vert_c, vpx_scaled_2d_c, vpx_scaled_avg_2d_c, 0);
vpx_scaled_avg_vert_c, vpx_scaled_2d_neon, vpx_scaled_avg_2d_c, 0);
const ConvolveParam kArrayConvolve_neon[] = { ALL_SIZES(convolve8_neon) };
#endif // CONFIG_VP9_HIGHBITDEPTH
......
......@@ -47,6 +47,28 @@ static INLINE void load_u8_8x8(const uint8_t *s, const ptrdiff_t p,
*s7 = vld1_u8(s);
}
static INLINE void load_u8_16x8(const uint8_t *s, const ptrdiff_t p,
uint8x16_t *const s0, uint8x16_t *const s1,
uint8x16_t *const s2, uint8x16_t *const s3,
uint8x16_t *const s4, uint8x16_t *const s5,
uint8x16_t *const s6, uint8x16_t *const s7) {
*s0 = vld1q_u8(s);
s += p;
*s1 = vld1q_u8(s);
s += p;
*s2 = vld1q_u8(s);
s += p;
*s3 = vld1q_u8(s);
s += p;
*s4 = vld1q_u8(s);
s += p;
*s5 = vld1q_u8(s);
s += p;
*s6 = vld1q_u8(s);
s += p;
*s7 = vld1q_u8(s);
}
static INLINE int16x4_t convolve8_4(const int16x4_t s0, const int16x4_t s1,
const int16x4_t s2, const int16x4_t s3,
const int16x4_t s4, const int16x4_t s5,
......
This diff is collapsed.
......@@ -106,6 +106,7 @@ DSP_SRCS-$(HAVE_NEON) += arm/highbd_vpx_convolve_neon.c
endif
DSP_SRCS-$(HAVE_SSE2) += x86/vpx_convolve_copy_sse2.asm
DSP_SRCS-$(HAVE_NEON) += arm/vpx_scaled_convolve8_neon.c
ifeq ($(HAVE_NEON_ASM),yes)
DSP_SRCS-yes += arm/vpx_convolve_copy_neon_asm$(ASM)
......
......@@ -365,7 +365,7 @@ add_proto qw/void vpx_convolve8_avg_vert/, "const uint8_t *src, ptrdiff_t src_st
specialize qw/vpx_convolve8_avg_vert sse2 ssse3 neon dspr2 msa vsx/;
add_proto qw/void vpx_scaled_2d/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h";
specialize qw/vpx_scaled_2d ssse3/;
specialize qw/vpx_scaled_2d ssse3 neon/;
add_proto qw/void vpx_scaled_horiz/, "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const InterpKernel *filter, int x0_q4, int x_step_q4, int y0_q4, int y_step_q4, int w, int h";
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment