[FFmpeg-devel] [PATCH] libavcodec/aarch64/hevcdsp_idct_neon.S: Also port add_residual functions.
Reimar.Doeffinger at gmx.de
Reimar.Doeffinger at gmx.de
Sun Jan 10 11:18:30 EET 2021
From: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
Speedup is fairly small, around 1.5%, but these are fairly simple.
---
libavcodec/aarch64/hevcdsp_idct_neon.S | 190 ++++++++++++++++++++++
libavcodec/aarch64/hevcdsp_init_aarch64.c | 24 +++
2 files changed, 214 insertions(+)
diff --git a/libavcodec/aarch64/hevcdsp_idct_neon.S b/libavcodec/aarch64/hevcdsp_idct_neon.S
index 9f67e45..edd03a0 100644
--- a/libavcodec/aarch64/hevcdsp_idct_neon.S
+++ b/libavcodec/aarch64/hevcdsp_idct_neon.S
@@ -36,6 +36,196 @@ const trans, align=4
.short 31, 22, 13, 4
endconst
+.macro clip10 in1, in2, c1, c2
+ smax \in1, \in1, \c1
+ smax \in2, \in2, \c1
+ smin \in1, \in1, \c2
+ smin \in2, \in2, \c2
+.endm
+
+function ff_hevc_add_residual_4x4_8_neon, export=1
+ ld1 {v0.8H-v1.8H}, [x1]
+ ld1 {v2.S}[0], [x0], x2
+ ld1 {v2.S}[1], [x0], x2
+ ld1 {v2.S}[2], [x0], x2
+ ld1 {v2.S}[3], [x0], x2
+ sub x0, x0, x2, lsl #2
+ uxtl v8.8H, v2.8B
+ uxtl2 v9.8H, v2.16B
+ sqadd v0.8H, v0.8H, v8.8H
+ sqadd v1.8H, v1.8H, v9.8H
+ sqxtun v0.8B, v0.8H
+ sqxtun2 v0.16B, v1.8H
+ st1 {v0.S}[0], [x0], x2
+ st1 {v0.S}[1], [x0], x2
+ st1 {v0.S}[2], [x0], x2
+ st1 {v0.S}[3], [x0], x2
+ ret
+endfunc
+
+function ff_hevc_add_residual_4x4_10_neon, export=1
+ mov x12, x0
+ ld1 {v0.8H-v1.8H}, [x1]
+ ld1 {v2.D}[0], [x12], x2
+ ld1 {v2.D}[1], [x12], x2
+ ld1 {v3.D}[0], [x12], x2
+ sqadd v0.8H, v0.8H, v2.8H
+ ld1 {V3.D}[1], [x12], x2
+ movi v4.8H, #0
+ sqadd v1.8H, v1.8H, v3.8H
+ mvni v5.8H, #0xFC, LSL #8 // movi #0x3FF
+ clip10 v0.8H, v1.8H, v4.8H, v5.8H
+ st1 {v0.D}[0], [x0], x2
+ st1 {v0.D}[1], [x0], x2
+ st1 {v1.D}[0], [x0], x2
+ st1 {v1.D}[1], [x0], x2
+ ret
+endfunc
+
+function ff_hevc_add_residual_8x8_8_neon, export=1
+ add x12, x0, x2
+ add x2, x2, x2
+ mov x3, #8
+1: subs x3, x3, #2
+ ld1 {v2.D}[0], [x0]
+ ld1 {v2.D}[1], [x12]
+ uxtl v3.8H, v2.8B
+ ld1 {v0.8H-v1.8H}, [x1], #32
+ uxtl2 v2.8H, v2.16B
+ sqadd v0.8H, v0.8H, v3.8H
+ sqadd v1.8H, v1.8H, v2.8H
+ sqxtun v0.8B, v0.8H
+ sqxtun2 v0.16B, v1.8H
+ st1 {v0.D}[0], [x0], x2
+ st1 {v0.D}[1], [x12], x2
+ bne 1b
+ ret
+endfunc
+
+function ff_hevc_add_residual_8x8_10_neon, export=1
+ add x12, x0, x2
+ add x2, x2, x2
+ mov x3, #8
+ movi v4.8H, #0
+ mvni v5.8H, #0xFC, LSL #8 // movi #0x3FF
+1: subs x3, x3, #2
+ ld1 {v0.8H-v1.8H}, [x1], #32
+ ld1 {v2.8H}, [x0]
+ sqadd v0.8H, v0.8H, v2.8H
+ ld1 {v3.8H}, [x12]
+ sqadd v1.8H, v1.8H, v3.8H
+ clip10 v0.8H, v1.8H, v4.8H, v5.8H
+ st1 {v0.8H}, [x0], x2
+ st1 {v1.8H}, [x12], x2
+ bne 1b
+ ret
+endfunc
+
+function ff_hevc_add_residual_16x16_8_neon, export=1
+ mov x3, #16
+ add x12, x0, x2
+ add x2, x2, x2
+1: subs x3, x3, #2
+ ld1 {v16.16B}, [x0]
+ ld1 {v0.8H-v3.8H}, [x1], #64
+ ld1 {v19.16B}, [x12]
+ uxtl v17.8H, v16.8B
+ uxtl2 v18.8H, v16.16B
+ uxtl v20.8H, v19.8B
+ uxtl2 v21.8H, v19.16B
+ sqadd v0.8H, v0.8H, v17.8H
+ sqadd v1.8H, v1.8H, v18.8H
+ sqadd v2.8H, v2.8H, v20.8H
+ sqadd v3.8H, v3.8H, v21.8H
+ sqxtun v0.8B, v0.8H
+ sqxtun2 v0.16B, v1.8H
+ sqxtun v1.8B, v2.8H
+ sqxtun2 v1.16B, v3.8H
+ st1 {v0.16B}, [x0], x2
+ st1 {v1.16B}, [x12], x2
+ bne 1b
+ ret
+endfunc
+
+function ff_hevc_add_residual_16x16_10_neon, export=1
+ mov x3, #16
+ movi v20.8H, #0
+ mvni v21.8H, #0xFC, LSL #8 // movi #0x3FF
+ add x12, x0, x2
+ add x2, x2, x2
+1: subs x3, x3, #2
+ ld1 {v16.8H-v17.8H}, [x0]
+ ld1 {v0.8H-v3.8H}, [x1], #64
+ sqadd v0.8H, v0.8H, v16.8H
+ ld1 {v18.8H-v19.8H}, [x12]
+ sqadd v1.8H, v1.8H, v17.8H
+ sqadd v2.8H, v2.8H, v18.8H
+ sqadd v3.8H, v3.8H, v19.8H
+ clip10 v0.8H, v1.8H, v20.8H, v21.8H
+ clip10 v2.8H, v3.8H, v20.8H, v21.8H
+ st1 {v0.8H-v1.8H}, [x0], x2
+ st1 {v2.8H-v3.8H}, [x12], x2
+ bne 1b
+ ret
+endfunc
+
+function ff_hevc_add_residual_32x32_8_neon, export=1
+ add x12, x0, x2
+ add x2, x2, x2
+ mov x3, #32
+1: subs x3, x3, #2
+ ld1 {v20.16B, v21.16B}, [x0]
+ uxtl v16.8H, v20.8B
+ uxtl2 v17.8H, v20.16B
+ ld1 {v22.16B, v23.16B}, [x12]
+ uxtl v18.8H, v21.8B
+ uxtl2 v19.8H, v21.16B
+ uxtl v20.8H, v22.8B
+ ld1 {v0.8H-v3.8H}, [x1], #64
+ ld1 {v4.8H-v7.8H}, [x1], #64
+ uxtl2 v21.8H, v22.16B
+ uxtl v22.8H, v23.8B
+ uxtl2 v23.8H, v23.16B
+ sqadd v0.8H, v0.8H, v16.8H
+ sqadd v1.8H, v1.8H, v17.8H
+ sqadd v2.8H, v2.8H, v18.8H
+ sqadd v3.8H, v3.8H, v19.8H
+ sqadd v4.8H, v4.8H, v20.8H
+ sqadd v5.8H, v5.8H, v21.8H
+ sqadd v6.8H, v6.8H, v22.8H
+ sqadd v7.8H, v7.8H, v23.8H
+ sqxtun v0.8B, v0.8H
+ sqxtun2 v0.16B, v1.8H
+ sqxtun v1.8B, v2.8H
+ sqxtun2 v1.16B, v3.8H
+ sqxtun v2.8B, v4.8H
+ sqxtun2 v2.16B, v5.8H
+ st1 {v0.16B, v1.16B}, [x0], x2
+ sqxtun v3.8B, v6.8H
+ sqxtun2 v3.16B, v7.8H
+ st1 {v2.16B, v3.16B}, [x12], x2
+ bne 1b
+ ret
+endfunc
+
+function ff_hevc_add_residual_32x32_10_neon, export=1
+ mov x3, #32
+ movi v20.8H, #0
+ mvni v21.8H, #0xFC, LSL #8 // movi #0x3FF
+1: subs x3, x3, #1
+ ld1 {v0.8H-v3.8H}, [x1], #64
+ ld1 {v16.8H-v19.8H}, [x0]
+ sqadd v0.8H, v0.8H, v16.8H
+ sqadd v1.8H, v1.8H, v17.8H
+ sqadd v2.8H, v2.8H, v18.8H
+ sqadd v3.8H, v3.8H, v19.8H
+ clip10 v0.8H, v1.8H, v20.8H, v21.8H
+ clip10 v2.8H, v3.8H, v20.8H, v21.8H
+ st1 {v0.8H-v3.8H}, [x0], x2
+ bne 1b
+ ret
+endfunc
+
.macro sum_sub out, in, c, op, p
.ifc \op, +
smlal\p \out, \in, \c
diff --git a/libavcodec/aarch64/hevcdsp_init_aarch64.c b/libavcodec/aarch64/hevcdsp_init_aarch64.c
index 19d9a7f..4c29daa 100644
--- a/libavcodec/aarch64/hevcdsp_init_aarch64.c
+++ b/libavcodec/aarch64/hevcdsp_init_aarch64.c
@@ -25,6 +25,22 @@
#include "libavutil/aarch64/cpu.h"
#include "libavcodec/hevcdsp.h"
+void ff_hevc_add_residual_4x4_8_neon(uint8_t *_dst, int16_t *coeffs,
+ ptrdiff_t stride);
+void ff_hevc_add_residual_4x4_10_neon(uint8_t *_dst, int16_t *coeffs,
+ ptrdiff_t stride);
+void ff_hevc_add_residual_8x8_8_neon(uint8_t *_dst, int16_t *coeffs,
+ ptrdiff_t stride);
+void ff_hevc_add_residual_8x8_10_neon(uint8_t *_dst, int16_t *coeffs,
+ ptrdiff_t stride);
+void ff_hevc_add_residual_16x16_8_neon(uint8_t *_dst, int16_t *coeffs,
+ ptrdiff_t stride);
+void ff_hevc_add_residual_16x16_10_neon(uint8_t *_dst, int16_t *coeffs,
+ ptrdiff_t stride);
+void ff_hevc_add_residual_32x32_8_neon(uint8_t *_dst, int16_t *coeffs,
+ ptrdiff_t stride);
+void ff_hevc_add_residual_32x32_10_neon(uint8_t *_dst, int16_t *coeffs,
+ ptrdiff_t stride);
void ff_hevc_idct_8x8_8_neon(int16_t *coeffs, int col_limit);
void ff_hevc_idct_8x8_10_neon(int16_t *coeffs, int col_limit);
void ff_hevc_idct_16x16_8_neon(int16_t *coeffs, int col_limit);
@@ -35,10 +51,18 @@ av_cold void ff_hevc_dsp_init_aarch64(HEVCDSPContext *c, const int bit_depth)
if (!have_neon(av_get_cpu_flags())) return;
if (bit_depth == 8) {
+ c->add_residual[0] = ff_hevc_add_residual_4x4_8_neon;
+ c->add_residual[1] = ff_hevc_add_residual_8x8_8_neon;
+ c->add_residual[2] = ff_hevc_add_residual_16x16_8_neon;
+ c->add_residual[3] = ff_hevc_add_residual_32x32_8_neon;
c->idct[1] = ff_hevc_idct_8x8_8_neon;
c->idct[2] = ff_hevc_idct_16x16_8_neon;
}
if (bit_depth == 10) {
+ c->add_residual[0] = ff_hevc_add_residual_4x4_10_neon;
+ c->add_residual[1] = ff_hevc_add_residual_8x8_10_neon;
+ c->add_residual[2] = ff_hevc_add_residual_16x16_10_neon;
+ c->add_residual[3] = ff_hevc_add_residual_32x32_10_neon;
c->idct[1] = ff_hevc_idct_8x8_10_neon;
c->idct[2] = ff_hevc_idct_16x16_10_neon;
}
--
2.24.3 (Apple Git-128)
More information about the ffmpeg-devel
mailing list