On Wed, Aug 14, 2024 at 7:21 PM Martin Storsjö <mar...@martin.st> wrote:
> On Fri, 9 Aug 2024, Ramiro Polla wrote:
> > checkasm --bench for Raspberry Pi 5 Model B Rev 1.0:
> > nv24_yuv420p_128_c: 423.0
> > nv24_yuv420p_128_neon: 115.7
> > nv24_yuv420p_1920_c: 5939.5
> > nv24_yuv420p_1920_neon: 1339.7
> > nv42_yuv420p_128_c: 423.2
> > nv42_yuv420p_128_neon: 115.7
> > nv42_yuv420p_1920_c: 5907.5
> > nv42_yuv420p_1920_neon: 1342.5
> > ---
> > libswscale/aarch64/Makefile | 1 +
> > libswscale/aarch64/swscale_unscaled.c | 30 +++++++++
> > libswscale/aarch64/swscale_unscaled_neon.S | 75 ++++++++++++++++++++++
> > 3 files changed, 106 insertions(+)
> > create mode 100644 libswscale/aarch64/swscale_unscaled_neon.S
>
> > diff --git a/libswscale/aarch64/swscale_unscaled_neon.S
> > b/libswscale/aarch64/swscale_unscaled_neon.S
> > new file mode 100644
> > index 0000000000..a206fda41f
> > --- /dev/null
> > +++ b/libswscale/aarch64/swscale_unscaled_neon.S
> > @@ -0,0 +1,75 @@
> > +/*
> > + * Copyright (c) 2024 Ramiro Polla
> > + *
> > + * This file is part of FFmpeg.
> > + *
> > + * FFmpeg is free software; you can redistribute it and/or
> > + * modify it under the terms of the GNU Lesser General Public
> > + * License as published by the Free Software Foundation; either
> > + * version 2.1 of the License, or (at your option) any later version.
> > + *
> > + * FFmpeg is distributed in the hope that it will be useful,
> > + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> > + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
> > + * Lesser General Public License for more details.
> > + *
> > + * You should have received a copy of the GNU Lesser General Public
> > + * License along with FFmpeg; if not, write to the Free Software
> > + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
> > 02110-1301 USA
> > + */
> > +
> > +#include "libavutil/aarch64/asm.S"
> > +
> > +function ff_nv24_to_yuv420p_chroma_neon, export=1
> > +// x0 uint8_t *dst1
> > +// x1 int dstStride1
> > +// x2 uint8_t *dst2
> > +// x3 int dstStride2
> > +// x4 const uint8_t *src
> > +// x5 int srcStride
> > +// w6 int w
> > +// w7 int h
> > +
> > + uxtw x1, w1
> > + uxtw x3, w3
> > + uxtw x5, w5
>
> You can often avoid the explicit uxtw instructions, if you can fold an
> uxtw attribute into the cases where the register is used. (If it's used
> often, it may be slightly more performant to do it upfront like this
> though, but often it can be omitted entirely.) And whenever you do an
> operation with a wN register as destination, the upper half of the
> register gets explicitly cleared, so these also may be avoided that way.
>
> > +
> > + add x9, x4, x5 // x9 = src + srcStride
> > + lsl w5, w5, #1 // srcStride *= 2
> > +
> > +1:
> > + mov w10, w6 // w10 = w
> > + mov x11, x4 // x11 = src1 (line 1)
> > + mov x12, x9 // x12 = src2 (line 2)
> > + mov x13, x0 // x13 = dst1 (dstU)
> > + mov x14, x2 // x14 = dst2 (dstV)
> > +
> > +2:
> > + ld2 { v0.16b, v1.16b }, [x11], #32 // v0 = U1, v1 = V1
> > + ld2 { v2.16b, v3.16b }, [x12], #32 // v2 = U2, v3 = V2
> > +
> > + uaddlp v0.8h, v0.16b // pairwise add U1
> > into v0
> > + uaddlp v1.8h, v1.16b // pairwise add V1
> > into v1
> > + uadalp v0.8h, v2.16b // pairwise add U2,
> > accumulate into v0
> > + uadalp v1.8h, v3.16b // pairwise add V2,
> > accumulate into v1
> > +
> > + shrn v0.8b, v0.8h, #2 // divide by 4
> > + shrn v1.8b, v1.8h, #2 // divide by 4
> > +
> > + st1 { v0.8b }, [x13], #8 // store U into dst1
> > + st1 { v1.8b }, [x14], #8 // store V into dst2
> > +
> > + subs w10, w10, #8
> > + b.gt 2b
> > +
> > + // next row
> > + add x4, x4, x5 // src1 += srcStride *
> > 2
> > + add x9, x9, x5 // src2 += srcStride *
> > 2
> > + add x0, x0, x1 // dst1 += dstStride1
> > + add x2, x2, x3 // dst2 += dstStride2
>
> It's often possible to avoid the extra step of moving the pointers back
> into the the x11/x12/x13/x14 registers, if you subtract the width from the
> stride at the start of the function. Then you don't need two separate
> registers for each pointer, and shortens dependency chain when moving on
> to the next line.
>
> If the width can be any uneven value, but we in practice write in
> increments of 8 pixels, you may need to align the width up to 8 before
> using it to decrement the stride that way though.
Thank you for the review. New patch attached.
From f6ea1edb0590c14e168fbce2ae42958220b6e778 Mon Sep 17 00:00:00 2001
From: Ramiro Polla <ramiro.po...@gmail.com>
Date: Wed, 7 Aug 2024 18:53:12 +0200
Subject: [PATCH v2 4/4] swscale/aarch64: add nv24/nv42 to yuv420p unscaled
converter
checkasm --bench for Raspberry Pi 5 Model B Rev 1.0:
nv24_yuv420p_128_c: 1320.2
nv24_yuv420p_128_neon: 709.5
nv24_yuv420p_1920_c: 12448.0
nv24_yuv420p_1920_neon: 2698.0
nv42_yuv420p_128_c: 1329.2
nv42_yuv420p_128_neon: 841.7
nv42_yuv420p_1920_c: 11967.5
nv42_yuv420p_1920_neon: 2866.5
---
libswscale/aarch64/Makefile | 1 +
libswscale/aarch64/swscale_unscaled.c | 30 ++++++++++
libswscale/aarch64/swscale_unscaled_neon.S | 70 ++++++++++++++++++++++
3 files changed, 101 insertions(+)
create mode 100644 libswscale/aarch64/swscale_unscaled_neon.S
diff --git a/libswscale/aarch64/Makefile b/libswscale/aarch64/Makefile
index 37ad960619..1de8c9c0d6 100644
--- a/libswscale/aarch64/Makefile
+++ b/libswscale/aarch64/Makefile
@@ -7,4 +7,5 @@ NEON-OBJS += aarch64/hscale.o \
aarch64/output.o \
aarch64/range_convert_neon.o \
aarch64/rgb2rgb_neon.o \
+ aarch64/swscale_unscaled_neon.o \
aarch64/yuv2rgb_neon.o \
diff --git a/libswscale/aarch64/swscale_unscaled.c b/libswscale/aarch64/swscale_unscaled.c
index b3093bbc9d..87bb011709 100644
--- a/libswscale/aarch64/swscale_unscaled.c
+++ b/libswscale/aarch64/swscale_unscaled.c
@@ -83,6 +83,31 @@ static int ifmt##_to_##ofmt##_neon_wrapper(SwsContext *c, const uint8_t *src[],
c->yuv2rgb_y_coeff); \
} \
+void ff_nv24_to_yuv420p_chroma_neon(uint8_t *dst1, int dstStride1,
+ uint8_t *dst2, int dstStride2,
+ const uint8_t *src, int srcStride,
+ int w, int h);
+
+static int nv24_to_yuv420p_neon_wrapper(SwsContext *c, const uint8_t *src[],
+ int srcStride[], int srcSliceY, int srcSliceH,
+ uint8_t *dst[], int dstStride[])
+{
+ uint8_t *dst1 = dst[1] + dstStride[1] * srcSliceY / 2;
+ uint8_t *dst2 = dst[2] + dstStride[2] * srcSliceY / 2;
+
+ ff_copyPlane(src[0], srcStride[0], srcSliceY, srcSliceH, c->srcW,
+ dst[0], dstStride[0]);
+
+ if (c->srcFormat == AV_PIX_FMT_NV24)
+ ff_nv24_to_yuv420p_chroma_neon(dst1, dstStride[1], dst2, dstStride[2],
+ src[1], srcStride[1], c->srcW / 2, srcSliceH);
+ else
+ ff_nv24_to_yuv420p_chroma_neon(dst2, dstStride[2], dst1, dstStride[1],
+ src[1], srcStride[1], c->srcW / 2, srcSliceH);
+
+ return srcSliceH;
+}
+
#define DECLARE_FF_NVX_TO_ALL_RGBX_FUNCS(nvx) \
DECLARE_FF_NVX_TO_RGBX_FUNCS(nvx, argb) \
DECLARE_FF_NVX_TO_RGBX_FUNCS(nvx, rgba) \
@@ -119,6 +144,11 @@ static void get_unscaled_swscale_neon(SwsContext *c) {
SET_FF_NVX_TO_ALL_RGBX_FUNC(nv21, NV21, accurate_rnd);
SET_FF_NVX_TO_ALL_RGBX_FUNC(yuv420p, YUV420P, accurate_rnd);
SET_FF_NVX_TO_ALL_RGBX_FUNC(yuv422p, YUV422P, accurate_rnd);
+
+ if (c->dstFormat == AV_PIX_FMT_YUV420P &&
+ (c->srcFormat == AV_PIX_FMT_NV24 || c->srcFormat == AV_PIX_FMT_NV42) &&
+ !(c->srcH & 1) && !(c->srcW & 15) && !accurate_rnd)
+ c->convert_unscaled = nv24_to_yuv420p_neon_wrapper;
}
void ff_get_unscaled_swscale_aarch64(SwsContext *c)
diff --git a/libswscale/aarch64/swscale_unscaled_neon.S b/libswscale/aarch64/swscale_unscaled_neon.S
new file mode 100644
index 0000000000..7f1890f58a
--- /dev/null
+++ b/libswscale/aarch64/swscale_unscaled_neon.S
@@ -0,0 +1,70 @@
+/*
+ * Copyright (c) 2024 Ramiro Polla
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/aarch64/asm.S"
+
+function ff_nv24_to_yuv420p_chroma_neon, export=1
+// x0 uint8_t *dst1
+// x1 int dstStride1
+// x2 uint8_t *dst2
+// x3 int dstStride2
+// x4 const uint8_t *src
+// x5 int srcStride
+// w6 int w
+// w7 int h
+
+ add x9, x4, w5, sxtw // x9 = src + srcStride
+ lsl w5, w5, #1 // srcStride *= 2
+ sub w5, w5, w6, lsl #2 // srcPadding = (2 * srcStride) - (4 * w)
+ sub w1, w1, w6 // dstPadding1 = dstStride1 - w
+ sub w3, w3, w6 // dstPadding2 = dstStride2 - w
+
+1:
+ mov w10, w6 // w10 = w
+
+2:
+ ld2 { v0.16b, v1.16b }, [x4], #32 // v0 = U1, v1 = V1
+ ld2 { v2.16b, v3.16b }, [x9], #32 // v2 = U2, v3 = V2
+
+ uaddlp v0.8h, v0.16b // pairwise add U1 into v0
+ uaddlp v1.8h, v1.16b // pairwise add V1 into v1
+ uadalp v0.8h, v2.16b // pairwise add U2, accumulate into v0
+ uadalp v1.8h, v3.16b // pairwise add V2, accumulate into v1
+
+ shrn v0.8b, v0.8h, #2 // divide by 4
+ shrn v1.8b, v1.8h, #2 // divide by 4
+
+ st1 { v0.8b }, [x0], #8 // store U into dst1
+ st1 { v1.8b }, [x2], #8 // store V into dst2
+
+ subs w10, w10, #8
+ b.gt 2b
+
+ // next row
+ add x4, x4, x5, sxtw // src1 += srcPadding
+ add x9, x9, x5, sxtw // src2 += srcPadding
+ add x0, x0, x1, sxtw // dst1 += dstPadding1
+ add x2, x2, x3, sxtw // dst2 += dstPadding2
+
+ subs w7, w7, #2
+ b.gt 1b
+
+ ret
+endfunc
--
2.39.2
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
