Apr 30, 2024, 20:05 by nuomi2...@gmail.com:
> On Mon, Apr 29, 2024 at 11:38 PM Lynne <d...@lynne.ee> wrote:
>
>> Apr 29, 2024, 17:25 by toq...@outlook.com:
>>
>> > From: Wu Jianhua <toq...@outlook.com>
>> >
>> > Signed-off-by: Wu Jianhua <toq...@outlook.com>
>> > ---
>> > tests/checkasm/Makefile | 2 +-
>> > tests/checkasm/checkasm.c | 3 +-
>> > tests/checkasm/checkasm.h | 1 +
>> > tests/checkasm/vvc_alf.c | 133 ++++++++++++++++++++++++++++++++++++++
>> > 4 files changed, 137 insertions(+), 2 deletions(-)
>> > create mode 100644 tests/checkasm/vvc_alf.c
>> >
>> > diff --git a/tests/checkasm/Makefile b/tests/checkasm/Makefile
>> > index 2673e1d098..5a3e3985c4 100644
>> > --- a/tests/checkasm/Makefile
>> > +++ b/tests/checkasm/Makefile
>> > @@ -41,7 +41,7 @@ AVCODECOBJS-$(CONFIG_V210_DECODER) += v210dec.o
>> > AVCODECOBJS-$(CONFIG_V210_ENCODER) += v210enc.o
>> > AVCODECOBJS-$(CONFIG_VORBIS_DECODER) += vorbisdsp.o
>> > AVCODECOBJS-$(CONFIG_VP9_DECODER) += vp9dsp.o
>> > -AVCODECOBJS-$(CONFIG_VVC_DECODER) += vvc_mc.o
>> > +AVCODECOBJS-$(CONFIG_VVC_DECODER) += vvc_alf.o vvc_mc.o
>> >
>> > CHECKASMOBJS-$(CONFIG_AVCODEC) += $(AVCODECOBJS-yes)
>> >
>> > diff --git a/tests/checkasm/checkasm.c b/tests/checkasm/checkasm.c
>> > index 8be6cb0f55..8b2bf2827b 100644
>> > --- a/tests/checkasm/checkasm.c
>> > +++ b/tests/checkasm/checkasm.c
>> > @@ -198,7 +198,8 @@ static const struct {
>> > { "vorbisdsp", checkasm_check_vorbisdsp },
>> > #endif
>> > #if CONFIG_VVC_DECODER
>> > - { "vvc_mc", checkasm_check_vvc_mc },
>> > + { "vvc_alf", checkasm_check_vvc_alf },
>> > + { "vvc_mc", checkasm_check_vvc_mc },
>> > #endif
>> > #endif
>> > #if CONFIG_AVFILTER
>> > diff --git a/tests/checkasm/checkasm.h b/tests/checkasm/checkasm.h
>> > index f90920dee7..c6a5cf42dd 100644
>> > --- a/tests/checkasm/checkasm.h
>> > +++ b/tests/checkasm/checkasm.h
>> > @@ -132,6 +132,7 @@ void checkasm_check_vp8dsp(void);
>> > void checkasm_check_vp9dsp(void);
>> > void checkasm_check_videodsp(void);
>> > void checkasm_check_vorbisdsp(void);
>> > +void checkasm_check_vvc_alf(void);
>> > void checkasm_check_vvc_mc(void);
>> >
>> > struct CheckasmPerf;
>> > diff --git a/tests/checkasm/vvc_alf.c b/tests/checkasm/vvc_alf.c
>> > new file mode 100644
>> > index 0000000000..10469e1528
>> > --- /dev/null
>> > +++ b/tests/checkasm/vvc_alf.c
>> > @@ -0,0 +1,133 @@
>> > +/*
>> > + * Copyright (c) 2023-2024 Nuo Mi <nuomi2...@gmail.com>
>> > + * Copyright (c) 2023-2024 Wu Jianhua <toq...@outlook.com>
>> > + *
>> > + * This file is part of FFmpeg.
>> > + *
>> > + * FFmpeg is free software; you can redistribute it and/or modify
>> > + * it under the terms of the GNU General Public License as published by
>> > + * the Free Software Foundation; either version 2 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 General Public License for more details.
>> > + *
>> > + * You should have received a copy of the GNU 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 <string.h>
>> > +
>> > +#include "checkasm.h"
>> > +#include "libavcodec/vvc/ctu.h"
>> > +#include "libavcodec/vvc/data.h"
>> > +#include "libavcodec/vvc/dsp.h"
>> > +
>> > +#include "libavutil/common.h"
>> > +#include "libavutil/intreadwrite.h"
>> > +#include "libavutil/mem_internal.h"
>> > +
>> > +static const uint32_t pixel_mask[3] = { 0xffffffff, 0x03ff03ff,
>> 0x0fff0fff };
>> > +
>> > +#define SIZEOF_PIXEL ((bit_depth + 7) / 8)
>> > +#define SRC_PIXEL_STRIDE (MAX_CTU_SIZE + 2 * ALF_PADDING_SIZE)
>> > +#define DST_PIXEL_STRIDE (SRC_PIXEL_STRIDE + 4)
>> > +#define SRC_BUF_SIZE (SRC_PIXEL_STRIDE * (MAX_CTU_SIZE + 3 * 2) * 2)
>> //+3 * 2 for top and bottom row, *2 for high bit depth
>> > +#define DST_BUF_SIZE (DST_PIXEL_STRIDE * (MAX_CTU_SIZE + 3 * 2) * 2)
>> > +#define LUMA_PARAMS_SIZE (MAX_CTU_SIZE * MAX_CTU_SIZE / ALF_BLOCK_SIZE
>> / ALF_BLOCK_SIZE * ALF_NUM_COEFF_LUMA)
>> > +
>> > +#define randomize_buffers(buf0, buf1, size) \
>> > + do { \
>> > + uint32_t mask = pixel_mask[(bit_depth - 8) >> 1]; \
>> > + int k; \
>> > + for (k = 0; k < size; k += 4) { \
>> > + uint32_t r = rnd() & mask; \
>> > + AV_WN32A(buf0 + k, r); \
>> > + AV_WN32A(buf1 + k, r); \
>> > + } \
>> > + } while (0)
>> > +
>> > +#define randomize_buffers2(buf, size, filter) \
>> > + do { \
>> > + int k; \
>> > + if (filter) { \
>> > + for (k = 0; k < size; k++) { \
>> > + int8_t r = rnd(); \
>> > + buf[k] = r; \
>> > + } \
>> > + } else { \
>> > + for (k = 0; k < size; k++) { \
>> > + int r = rnd() % FF_ARRAY_ELEMS(clip_set); \
>> > + buf[k] = clip_set[r]; \
>> > + } \
>> > + } \
>> > + } while (0)
>> > +
>> > +static void check_alf_filter(VVCDSPContext *c, const int bit_depth)
>> > +{
>> > + LOCAL_ALIGNED_32(uint8_t, dst0, [DST_BUF_SIZE]);
>> > + LOCAL_ALIGNED_32(uint8_t, dst1, [DST_BUF_SIZE]);
>> > + LOCAL_ALIGNED_32(uint8_t, src0, [SRC_BUF_SIZE]);
>> > + LOCAL_ALIGNED_32(uint8_t, src1, [SRC_BUF_SIZE]);
>> > + int16_t filter[LUMA_PARAMS_SIZE];
>> > + int16_t clip[LUMA_PARAMS_SIZE];
>> > +
>> > + const int16_t clip_set[] = {
>> > + 1 << bit_depth, 1 << (bit_depth - 3), 1 << (bit_depth - 5), 1
>> << (bit_depth - 7)
>> > + };
>> > +
>> > + ptrdiff_t src_stride = SRC_PIXEL_STRIDE * SIZEOF_PIXEL;
>> > + ptrdiff_t dst_stride = DST_PIXEL_STRIDE * SIZEOF_PIXEL;
>> > + int offset = (3 * SRC_PIXEL_STRIDE + 3) * SIZEOF_PIXEL;
>> > +
>> > + declare_func_emms(AV_CPU_FLAG_AVX2, void, uint8_t *dst, ptrdiff_t
>> dst_stride, const uint8_t *src, ptrdiff_t src_stride,
>> > + int width, int height, const int16_t *filter, const int16_t
>> *clip, const int vb_pos);
>> > +
>> > + randomize_buffers(src0, src1, SRC_BUF_SIZE);
>> > + randomize_buffers2(filter, LUMA_PARAMS_SIZE, 1);
>> > + randomize_buffers2(clip, LUMA_PARAMS_SIZE, 0);
>> > +
>> > + for (int h = 4; h <= MAX_CTU_SIZE; h += 4) {
>> > + for (int w = 4; w <= MAX_CTU_SIZE; w += 4) {
>> >
>>
>> That's an excessive amount of tests, isn't it?
>>
> yeah, maybe we can test squares only.
>
Actually nevermind, disregard this, we test similar dimensions
for vp9's asm too, and I think dav1d does the same.
Gives us more data and will catch more failures.
_______________________________________________
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".
