[FFmpeg-devel] [PATCH] v210enc: Add x86 SIMD
From: Kieran Kunhya
Signed-off-by: Michael Niedermayer
---
libavcodec/v210enc.c | 78 +
libavcodec/v210enc.h | 31
libavcodec/x86/Makefile |2 ++
libavcodec/x86/v210enc.asm| 75 +++
libavcodec/x86/v210enc_init.c | 31
5 files changed, 194 insertions(+), 23 deletions(-)
create mode 100644 libavcodec/v210enc.h
create mode 100644 libavcodec/x86/v210enc.asm
create mode 100644 libavcodec/x86/v210enc_init.c
diff --git a/libavcodec/v210enc.c b/libavcodec/v210enc.c
index 1e53bdb..f4fc1fe 100644
--- a/libavcodec/v210enc.c
+++ b/libavcodec/v210enc.c
@@ -24,9 +24,37 @@
#include "avcodec.h"
#include "bytestream.h"
#include "internal.h"
+#include "v210enc.h"
+
+#define CLIP(v) av_clip(v, 4, 1019)
+
+#define WRITE_PIXELS(a, b, c) \
+do {\
+val = CLIP(*a++); \
+val |= (CLIP(*b++) << 10) | \
+ (CLIP(*c++) << 20); \
+AV_WL32(dst, val); \
+dst += 4; \
+} while (0)
+
+static void v210_planar_pack_c(const uint16_t *y, const uint16_t *u,
+ const uint16_t *v, uint8_t *dst, ptrdiff_t
width)
+{
+uint32_t val;
+int i;
+
+for( i = 0; i < width-5; i += 6 ){
+WRITE_PIXELS(u, y, v);
+WRITE_PIXELS(y, u, y);
+WRITE_PIXELS(v, y, u);
+WRITE_PIXELS(y, v, y);
+}
+}
static av_cold int encode_init(AVCodecContext *avctx)
{
+V210EncContext *s = avctx->priv_data;
+
if (avctx->width & 1) {
av_log(avctx, AV_LOG_ERROR, "v210 needs even width\n");
return AVERROR(EINVAL);
@@ -42,12 +70,19 @@ static av_cold int encode_init(AVCodecContext *avctx)
avctx->coded_frame->pict_type = AV_PICTURE_TYPE_I;
+s->pack_line= v210_planar_pack_c;
+
+if (HAVE_MMX)
+v210enc_x86_init(s);
+
return 0;
}
static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
const AVFrame *pic, int *got_packet)
{
+V210EncContext *s = avctx->priv_data;
+
int aligned_width = ((avctx->width + 47) / 48) * 48;
int stride = aligned_width * 8 / 3;
int line_padding = stride - ((avctx->width * 8 + 11) / 12) * 4;
@@ -55,47 +90,43 @@ static int encode_frame(AVCodecContext *avctx, AVPacket
*pkt,
const uint16_t *y = (const uint16_t*)pic->data[0];
const uint16_t *u = (const uint16_t*)pic->data[1];
const uint16_t *v = (const uint16_t*)pic->data[2];
-PutByteContext p;
+uint8_t *dst;
if ((ret = ff_alloc_packet2(avctx, pkt, avctx->height * stride)) < 0)
return ret;
-bytestream2_init_writer(&p, pkt->data, pkt->size);
-
-#define CLIP(v) av_clip(v, 4, 1019)
-
-#define WRITE_PIXELS(a, b, c) \
-do {\
-val = CLIP(*a++); \
-val |= (CLIP(*b++) << 10) | \
- (CLIP(*c++) << 20); \
-bytestream2_put_le32u(&p, val); \
-} while (0)
+dst = pkt->data;
for (h = 0; h < avctx->height; h++) {
uint32_t val;
-for (w = 0; w < avctx->width - 5; w += 6) {
-WRITE_PIXELS(u, y, v);
-WRITE_PIXELS(y, u, y);
-WRITE_PIXELS(v, y, u);
-WRITE_PIXELS(y, v, y);
-}
+w = (avctx->width / 6) * 6;
+s->pack_line(y, u, v, dst, w);
+
+y += w;
+u += w >> 1;
+v += w >> 1;
+dst += (w / 6) * 16;
if (w < avctx->width - 1) {
WRITE_PIXELS(u, y, v);
val = CLIP(*y++);
-if (w == avctx->width - 2)
-bytestream2_put_le32u(&p, val);
+if (w == avctx->width - 2) {
+AV_WL32(dst, val);
+dst += 4;
+}
if (w < avctx->width - 3) {
val |= (CLIP(*u++) << 10) | (CLIP(*y++) << 20);
-bytestream2_put_le32u(&p, val);
+AV_WL32(dst, val);
+dst += 4;
val = CLIP(*v++) | (CLIP(*y++) << 10);
-bytestream2_put_le32u(&p, val);
+AV_WL32(dst, val);
+dst += 4;
}
}
-bytestream2_set_buffer(&p, 0, line_padding);
+memset(dst, 0, line_padding);
+dst += line_padding;
y += pic->linesize[0] / 2 - avctx->width;
u += pic->linesize[1] / 2 - avctx->width / 2;
@@ -119,6 +150,7 @@ AVCodec ff_v210_encoder = {
.long_name = NULL_IF_CONFIG_SMALL("Uncompressed 4:2:2 10-bit"),
.type = AVMEDIA_TYPE_VIDEO,
.id = AV_CODEC_ID_V210,
+.priv_data_size = sizeof(V210EncContext),
.init = encode_init,
.encode2= encode_frame,
.close = encode_close,
diff --git a/libavcodec/v210enc.h b/libavcodec/v210enc.h
new file m
Re: [FFmpeg-devel] [PATCH] v210enc: Add x86 SIMD
Kieran Kunhya obe.tv> writes: > --- a/libavcodec/x86/Makefile > +++ b/libavcodec/x86/Makefile > -47,6 +47,7 >x86/rv40dsp_init.o > OBJS-$(CONFIG_SVQ1_ENCODER)+= x86/svq1enc.o > OBJS-$(CONFIG_TRUEHD_DECODER) += x86/mlpdsp.o > +OBJS-$(CONFIG_V210_ENCODER)+= x86/v210enc_init.o This does not apply here, please update to current git head. Carl Eugen ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
[FFmpeg-devel] [PATCH] v210enc: Add x86 SIMD
---
libavcodec/v210enc.c | 78 ++-
libavcodec/v210enc.h | 31 +
libavcodec/x86/Makefile | 2 ++
libavcodec/x86/v210enc.asm| 76 +
libavcodec/x86/v210enc_init.c | 31 +
5 files changed, 195 insertions(+), 23 deletions(-)
create mode 100644 libavcodec/v210enc.h
create mode 100644 libavcodec/x86/v210enc.asm
create mode 100644 libavcodec/x86/v210enc_init.c
diff --git a/libavcodec/v210enc.c b/libavcodec/v210enc.c
index ef0d6ab..4a6bdfc 100644
--- a/libavcodec/v210enc.c
+++ b/libavcodec/v210enc.c
@@ -24,9 +24,37 @@
#include "avcodec.h"
#include "bytestream.h"
#include "internal.h"
+#include "v210enc.h"
+
+#define CLIP(v) av_clip(v, 4, 1019)
+
+#define WRITE_PIXELS(a, b, c) \
+do {\
+val = CLIP(*a++); \
+val |= (CLIP(*b++) << 10) | \
+ (CLIP(*c++) << 20); \
+AV_WL32(dst, val); \
+dst += 4; \
+} while (0)
+
+static void v210_planar_pack_c(const uint16_t *y, const uint16_t *u,
+ const uint16_t *v, uint8_t *dst, ptrdiff_t
width)
+{
+uint32_t val;
+int i;
+
+for( i = 0; i < width-5; i += 6 ){
+WRITE_PIXELS(u, y, v);
+WRITE_PIXELS(y, u, y);
+WRITE_PIXELS(v, y, u);
+WRITE_PIXELS(y, v, y);
+}
+}
static av_cold int encode_init(AVCodecContext *avctx)
{
+V210EncContext *s = avctx->priv_data;
+
if (avctx->width & 1) {
av_log(avctx, AV_LOG_ERROR, "v210 needs even width\n");
return AVERROR(EINVAL);
@@ -42,12 +70,19 @@ static av_cold int encode_init(AVCodecContext *avctx)
avctx->coded_frame->pict_type = AV_PICTURE_TYPE_I;
+s->pack_line= v210_planar_pack_c;
+
+if (HAVE_MMX)
+v210enc_x86_init(s);
+
return 0;
}
static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
const AVFrame *pic, int *got_packet)
{
+V210EncContext *s = avctx->priv_data;
+
int aligned_width = ((avctx->width + 47) / 48) * 48;
int stride = aligned_width * 8 / 3;
int line_padding = stride - ((avctx->width * 8 + 11) / 12) * 4;
@@ -55,49 +90,45 @@ static int encode_frame(AVCodecContext *avctx, AVPacket
*pkt,
const uint16_t *y = (const uint16_t*)pic->data[0];
const uint16_t *u = (const uint16_t*)pic->data[1];
const uint16_t *v = (const uint16_t*)pic->data[2];
-PutByteContext p;
+uint8_t *dst;
if ((ret = ff_alloc_packet(pkt, avctx->height * stride)) < 0) {
av_log(avctx, AV_LOG_ERROR, "Error getting output packet.\n");
return ret;
}
-bytestream2_init_writer(&p, pkt->data, pkt->size);
-
-#define CLIP(v) av_clip(v, 4, 1019)
-
-#define WRITE_PIXELS(a, b, c) \
-do {\
-val = CLIP(*a++); \
-val |= (CLIP(*b++) << 10) | \
- (CLIP(*c++) << 20); \
-bytestream2_put_le32u(&p, val); \
-} while (0)
+dst = pkt->data;
for (h = 0; h < avctx->height; h++) {
uint32_t val;
-for (w = 0; w < avctx->width - 5; w += 6) {
-WRITE_PIXELS(u, y, v);
-WRITE_PIXELS(y, u, y);
-WRITE_PIXELS(v, y, u);
-WRITE_PIXELS(y, v, y);
-}
+w = (avctx->width / 6) * 6;
+s->pack_line(y, u, v, dst, w);
+
+y += w;
+u += w >> 1;
+v += w >> 1;
+dst += (w / 6) * 16;
if (w < avctx->width - 1) {
WRITE_PIXELS(u, y, v);
val = CLIP(*y++);
-if (w == avctx->width - 2)
-bytestream2_put_le32u(&p, val);
+if (w == avctx->width - 2) {
+AV_WL32(dst, val);
+dst += 4;
+}
}
if (w < avctx->width - 3) {
val |= (CLIP(*u++) << 10) | (CLIP(*y++) << 20);
-bytestream2_put_le32u(&p, val);
+AV_WL32(dst, val);
+dst += 4;
val = CLIP(*v++) | (CLIP(*y++) << 10);
-bytestream2_put_le32u(&p, val);
+AV_WL32(dst, val);
+dst += 4;
}
-bytestream2_set_buffer(&p, 0, line_padding);
+memset(dst, 0, line_padding);
+dst += line_padding;
y += pic->linesize[0] / 2 - avctx->width;
u += pic->linesize[1] / 2 - avctx->width / 2;
@@ -121,6 +152,7 @@ AVCodec ff_v210_encoder = {
.long_name = NULL_IF_CONFIG_SMALL("Uncompressed 4:2:2 10-bit"),
.type = AVMEDIA_TYPE_VIDEO,
.id = AV_CODEC_ID_V210,
+.priv_data_size = sizeof(V210EncContext),
.init = encode_init,
.encode2= encode_frame,
.close = encode_close,
diff --git a/libavcodec/v210enc.h b/libavcodec/v210enc.h
new file mode 100644
index 000..b8b
[FFmpeg-devel] [PATCH] v210enc: Add x86 SIMD
---
libavcodec/v210enc.c| 78 ++---
libavcodec/v210enc.h| 31
libavcodec/x86/Makefile | 2 ++
3 files changed, 88 insertions(+), 23 deletions(-)
create mode 100644 libavcodec/v210enc.h
diff --git a/libavcodec/v210enc.c b/libavcodec/v210enc.c
index ef0d6ab..4a6bdfc 100644
--- a/libavcodec/v210enc.c
+++ b/libavcodec/v210enc.c
@@ -24,9 +24,37 @@
#include "avcodec.h"
#include "bytestream.h"
#include "internal.h"
+#include "v210enc.h"
+
+#define CLIP(v) av_clip(v, 4, 1019)
+
+#define WRITE_PIXELS(a, b, c) \
+do {\
+val = CLIP(*a++); \
+val |= (CLIP(*b++) << 10) | \
+ (CLIP(*c++) << 20); \
+AV_WL32(dst, val); \
+dst += 4; \
+} while (0)
+
+static void v210_planar_pack_c(const uint16_t *y, const uint16_t *u,
+ const uint16_t *v, uint8_t *dst, ptrdiff_t
width)
+{
+uint32_t val;
+int i;
+
+for( i = 0; i < width-5; i += 6 ){
+WRITE_PIXELS(u, y, v);
+WRITE_PIXELS(y, u, y);
+WRITE_PIXELS(v, y, u);
+WRITE_PIXELS(y, v, y);
+}
+}
static av_cold int encode_init(AVCodecContext *avctx)
{
+V210EncContext *s = avctx->priv_data;
+
if (avctx->width & 1) {
av_log(avctx, AV_LOG_ERROR, "v210 needs even width\n");
return AVERROR(EINVAL);
@@ -42,12 +70,19 @@ static av_cold int encode_init(AVCodecContext *avctx)
avctx->coded_frame->pict_type = AV_PICTURE_TYPE_I;
+s->pack_line= v210_planar_pack_c;
+
+if (HAVE_MMX)
+v210enc_x86_init(s);
+
return 0;
}
static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
const AVFrame *pic, int *got_packet)
{
+V210EncContext *s = avctx->priv_data;
+
int aligned_width = ((avctx->width + 47) / 48) * 48;
int stride = aligned_width * 8 / 3;
int line_padding = stride - ((avctx->width * 8 + 11) / 12) * 4;
@@ -55,49 +90,45 @@ static int encode_frame(AVCodecContext *avctx, AVPacket
*pkt,
const uint16_t *y = (const uint16_t*)pic->data[0];
const uint16_t *u = (const uint16_t*)pic->data[1];
const uint16_t *v = (const uint16_t*)pic->data[2];
-PutByteContext p;
+uint8_t *dst;
if ((ret = ff_alloc_packet(pkt, avctx->height * stride)) < 0) {
av_log(avctx, AV_LOG_ERROR, "Error getting output packet.\n");
return ret;
}
-bytestream2_init_writer(&p, pkt->data, pkt->size);
-
-#define CLIP(v) av_clip(v, 4, 1019)
-
-#define WRITE_PIXELS(a, b, c) \
-do {\
-val = CLIP(*a++); \
-val |= (CLIP(*b++) << 10) | \
- (CLIP(*c++) << 20); \
-bytestream2_put_le32u(&p, val); \
-} while (0)
+dst = pkt->data;
for (h = 0; h < avctx->height; h++) {
uint32_t val;
-for (w = 0; w < avctx->width - 5; w += 6) {
-WRITE_PIXELS(u, y, v);
-WRITE_PIXELS(y, u, y);
-WRITE_PIXELS(v, y, u);
-WRITE_PIXELS(y, v, y);
-}
+w = (avctx->width / 6) * 6;
+s->pack_line(y, u, v, dst, w);
+
+y += w;
+u += w >> 1;
+v += w >> 1;
+dst += (w / 6) * 16;
if (w < avctx->width - 1) {
WRITE_PIXELS(u, y, v);
val = CLIP(*y++);
-if (w == avctx->width - 2)
-bytestream2_put_le32u(&p, val);
+if (w == avctx->width - 2) {
+AV_WL32(dst, val);
+dst += 4;
+}
}
if (w < avctx->width - 3) {
val |= (CLIP(*u++) << 10) | (CLIP(*y++) << 20);
-bytestream2_put_le32u(&p, val);
+AV_WL32(dst, val);
+dst += 4;
val = CLIP(*v++) | (CLIP(*y++) << 10);
-bytestream2_put_le32u(&p, val);
+AV_WL32(dst, val);
+dst += 4;
}
-bytestream2_set_buffer(&p, 0, line_padding);
+memset(dst, 0, line_padding);
+dst += line_padding;
y += pic->linesize[0] / 2 - avctx->width;
u += pic->linesize[1] / 2 - avctx->width / 2;
@@ -121,6 +152,7 @@ AVCodec ff_v210_encoder = {
.long_name = NULL_IF_CONFIG_SMALL("Uncompressed 4:2:2 10-bit"),
.type = AVMEDIA_TYPE_VIDEO,
.id = AV_CODEC_ID_V210,
+.priv_data_size = sizeof(V210EncContext),
.init = encode_init,
.encode2= encode_frame,
.close = encode_close,
diff --git a/libavcodec/v210enc.h b/libavcodec/v210enc.h
new file mode 100644
index 000..b8b6143
--- /dev/null
+++ b/libavcodec/v210enc.h
@@ -0,0 +1,31 @@
+/*
+ * This file is part of Libav.
+ *
+ * Libav is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as
