virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-18 03:19:31 +02:00
Code Issues Releases Activity
0930b2dd1f
FFmpeg/tests/checkasm/h264pred.c

261 lines
9.8 KiB
C
Raw Normal View History

Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
/*
* Copyright (c) 2015 Henrik Gramner
*
Merge commit '8bc67ec2c0d2b5444d51a1bed1d50f0e10d92717' * commit '8bc67ec2c0d2b5444d51a1bed1d50f0e10d92717': Checkasm: assembly testing and benchmarking tool Merged-by: Michael Niedermayer <michael@niedermayer.cc>
2015-07-12 21:00:50 +02:00
* This file is part of FFmpeg.
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
*
Merge commit '8bc67ec2c0d2b5444d51a1bed1d50f0e10d92717' * commit '8bc67ec2c0d2b5444d51a1bed1d50f0e10d92717': Checkasm: assembly testing and benchmarking tool Merged-by: Michael Niedermayer <michael@niedermayer.cc>
2015-07-12 21:00:50 +02:00
* FFmpeg is free software; you can redistribute it and/or modify
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
* 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.
*
Merge commit '8bc67ec2c0d2b5444d51a1bed1d50f0e10d92717' * commit '8bc67ec2c0d2b5444d51a1bed1d50f0e10d92717': Checkasm: assembly testing and benchmarking tool Merged-by: Michael Niedermayer <michael@niedermayer.cc>
2015-07-12 21:00:50 +02:00
* FFmpeg is distributed in the hope that it will be useful,
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
* 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
Merge commit '8bc67ec2c0d2b5444d51a1bed1d50f0e10d92717' * commit '8bc67ec2c0d2b5444d51a1bed1d50f0e10d92717': Checkasm: assembly testing and benchmarking tool Merged-by: Michael Niedermayer <michael@niedermayer.cc>
2015-07-12 21:00:50 +02:00
* with FFmpeg; if not, write to the Free Software Foundation, Inc.,
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
*/
#include <string.h>
#include "checkasm.h"
#include "libavcodec/avcodec.h"
#include "libavcodec/h264pred.h"
#include "libavutil/common.h"
tests/checkasm/h264pred: Use LOCAL_ALIGNED_16() Fixes alignment issue and bus errors Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2015-07-23 00:35:18 +02:00
#include "libavutil/internal.h"
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
#include "libavutil/intreadwrite.h"
static const int codec_ids[4] = { AV_CODEC_ID_H264, AV_CODEC_ID_VP8, AV_CODEC_ID_RV40, AV_CODEC_ID_SVQ3 };
static const char * const pred4x4_modes[4][15] = {
{ /* H264 */
[VERT_PRED ] = "vertical",
[HOR_PRED ] = "horizontal",
[DC_PRED ] = "dc",
[DIAG_DOWN_LEFT_PRED ] = "down_left",
[DIAG_DOWN_RIGHT_PRED] = "down_right",
[VERT_RIGHT_PRED ] = "vertical_right",
[HOR_DOWN_PRED ] = "horizontal_right",
[VERT_LEFT_PRED ] = "vertical_left",
[HOR_UP_PRED ] = "horizontal_up",
[LEFT_DC_PRED ] = "left_dc",
[TOP_DC_PRED ] = "top_dc",
[DC_128_PRED ] = "dc_128",
},
{ /* VP8 */
[VERT_PRED ] = "vertical_vp8",
[HOR_PRED ] = "horizontal_vp8",
[VERT_LEFT_PRED] = "vertical_left_vp8",
[TM_VP8_PRED ] = "tm_vp8",
[DC_127_PRED ] = "dc_127_vp8",
[DC_129_PRED ] = "dc_129_vp8",
},
{ /* RV40 */
[DIAG_DOWN_LEFT_PRED ] = "down_left_rv40",
[VERT_LEFT_PRED ] = "vertical_left_rv40",
[HOR_UP_PRED ] = "horizontal_up_rv40",
[DIAG_DOWN_LEFT_PRED_RV40_NODOWN] = "down_left_nodown_rv40",
[HOR_UP_PRED_RV40_NODOWN ] = "horizontal_up_nodown_rv40",
[VERT_LEFT_PRED_RV40_NODOWN ] = "vertical_left_nodown_rv40",
},
{ /* SVQ3 */
[DIAG_DOWN_LEFT_PRED] = "down_left_svq3",
},
};
static const char * const pred8x8_modes[4][11] = {
{ /* H264 */
[DC_PRED8x8 ] = "dc",
[HOR_PRED8x8 ] = "horizontal",
[VERT_PRED8x8 ] = "vertical",
[PLANE_PRED8x8 ] = "plane",
[LEFT_DC_PRED8x8 ] = "left_dc",
[TOP_DC_PRED8x8 ] = "top_dc",
[DC_128_PRED8x8 ] = "dc_128",
[ALZHEIMER_DC_L0T_PRED8x8] = "mad_cow_dc_l0t",
[ALZHEIMER_DC_0LT_PRED8x8] = "mad_cow_dc_0lt",
[ALZHEIMER_DC_L00_PRED8x8] = "mad_cow_dc_l00",
[ALZHEIMER_DC_0L0_PRED8x8] = "mad_cow_dc_0l0",
},
{ /* VP8 */
[PLANE_PRED8x8 ] = "tm_vp8",
[DC_127_PRED8x8] = "dc_127_vp8",
[DC_129_PRED8x8] = "dc_129_vp8",
},
{ /* RV40 */
[DC_PRED8x8 ] = "dc_rv40",
[LEFT_DC_PRED8x8] = "left_dc_rv40",
[TOP_DC_PRED8x8 ] = "top_dc_rv40",
},
checkasm: remove empty array initializer list in h264pred test Fixes MSVC compilation.
2015-07-20 23:17:57 +02:00
/* nothing for SVQ3 */
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
};
static const char * const pred16x16_modes[4][9] = {
{ /* H264 */
[DC_PRED8x8 ] = "dc",
[HOR_PRED8x8 ] = "horizontal",
[VERT_PRED8x8 ] = "vertical",
[PLANE_PRED8x8 ] = "plane",
[LEFT_DC_PRED8x8] = "left_dc",
[TOP_DC_PRED8x8 ] = "top_dc",
[DC_128_PRED8x8 ] = "dc_128",
},
{ /* VP8 */
[PLANE_PRED8x8 ] = "tm_vp8",
[DC_127_PRED8x8] = "dc_127_vp8",
[DC_129_PRED8x8] = "dc_129_vp8",
},
{ /* RV40 */
[PLANE_PRED8x8] = "plane_rv40",
},
{ /* SVQ3 */
[PLANE_PRED8x8] = "plane_svq3",
},
};
static const uint32_t pixel_mask[3] = { 0xffffffff, 0x01ff01ff, 0x03ff03ff };
#define SIZEOF_PIXEL ((bit_depth + 7) / 8)
cosmetics: Reformat checkasm tests
2015-07-17 20:07:26 +02:00
#define BUF_SIZE (3 * 16 * 17)
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
cosmetics: Reformat checkasm tests
2015-07-17 20:07:26 +02:00
#define check_pred_func(func, name, mode_name) \
(mode_name && ((codec_ids[codec] == AV_CODEC_ID_H264) ? \
check_func(func, "pred%s_%s_%d", name, mode_name, bit_depth) : \
check_func(func, "pred%s_%s", name, mode_name)))
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
cosmetics: Reformat checkasm tests
2015-07-17 20:07:26 +02:00
#define randomize_buffers() \
do { \
uint32_t mask = pixel_mask[bit_depth - 8]; \
int i; \
for (i = 0; i < BUF_SIZE; i += 4) { \
uint32_t r = rnd() & mask; \
AV_WN32A(buf0 + i, r); \
AV_WN32A(buf1 + i, r); \
} \
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
} while (0)
cosmetics: Reformat checkasm tests
2015-07-17 20:07:26 +02:00
#define src0 (buf0 + 4 * 16) /* Offset to allow room for top and left */
#define src1 (buf1 + 4 * 16)
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
static void check_pred4x4(H264PredContext *h, uint8_t *buf0, uint8_t *buf1,
int codec, int chroma_format, int bit_depth)
{
if (chroma_format == 1) {
uint8_t *topright = buf0 + 2*16;
int pred_mode;
x86: checkasm: check for or handle missing cleanup after MMX instructions Not every asm routine is expected clear the MMX state after returning. It is however a requisite for testing floating point code in checkasm. Annotate functions requiring cleanup with declare_func_emms() and issue emms after the call. The remaining functions are checked for having a cleared MMX state after return.
2015-12-11 15:06:38 +02:00
declare_func_emms(AV_CPU_FLAG_MMX | AV_CPU_FLAG_MMXEXT, void, uint8_t *src, const uint8_t *topright, ptrdiff_t stride);
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
for (pred_mode = 0; pred_mode < 15; pred_mode++) {
if (check_pred_func(h->pred4x4[pred_mode], "4x4", pred4x4_modes[codec][pred_mode])) {
randomize_buffers();
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
call_ref(src0, topright, 12*SIZEOF_PIXEL);
call_new(src1, topright, 12*SIZEOF_PIXEL);
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
if (memcmp(buf0, buf1, BUF_SIZE))
fail();
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
bench_new(src1, topright, 12*SIZEOF_PIXEL);
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
}
}
}
}
static void check_pred8x8(H264PredContext *h, uint8_t *buf0, uint8_t *buf1,
int codec, int chroma_format, int bit_depth)
{
int pred_mode;
x86: checkasm: check for or handle missing cleanup after MMX instructions Not every asm routine is expected clear the MMX state after returning. It is however a requisite for testing floating point code in checkasm. Annotate functions requiring cleanup with declare_func_emms() and issue emms after the call. The remaining functions are checked for having a cleared MMX state after return.
2015-12-11 15:06:38 +02:00
declare_func_emms(AV_CPU_FLAG_MMX | AV_CPU_FLAG_MMXEXT, void, uint8_t *src, ptrdiff_t stride);
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
for (pred_mode = 0; pred_mode < 11; pred_mode++) {
if (check_pred_func(h->pred8x8[pred_mode], (chroma_format == 2) ? "8x16" : "8x8",
pred8x8_modes[codec][pred_mode])) {
randomize_buffers();
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
call_ref(src0, 24*SIZEOF_PIXEL);
call_new(src1, 24*SIZEOF_PIXEL);
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
if (memcmp(buf0, buf1, BUF_SIZE))
fail();
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
bench_new(src1, 24*SIZEOF_PIXEL);
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
}
}
}
static void check_pred16x16(H264PredContext *h, uint8_t *buf0, uint8_t *buf1,
int codec, int chroma_format, int bit_depth)
{
if (chroma_format == 1) {
int pred_mode;
x86: checkasm: check for or handle missing cleanup after MMX instructions Not every asm routine is expected clear the MMX state after returning. It is however a requisite for testing floating point code in checkasm. Annotate functions requiring cleanup with declare_func_emms() and issue emms after the call. The remaining functions are checked for having a cleared MMX state after return.
2015-12-11 15:06:38 +02:00
declare_func_emms(AV_CPU_FLAG_MMX | AV_CPU_FLAG_MMXEXT, void, uint8_t *src, ptrdiff_t stride);
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
for (pred_mode = 0; pred_mode < 9; pred_mode++) {
if (check_pred_func(h->pred16x16[pred_mode], "16x16", pred16x16_modes[codec][pred_mode])) {
randomize_buffers();
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
call_ref(src0, 48);
call_new(src1, 48);
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
if (memcmp(buf0, buf1, BUF_SIZE))
fail();
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
bench_new(src1, 48);
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
}
}
}
}
static void check_pred8x8l(H264PredContext *h, uint8_t *buf0, uint8_t *buf1,
int codec, int chroma_format, int bit_depth)
{
if (chroma_format == 1 && codec_ids[codec] == AV_CODEC_ID_H264) {
int pred_mode;
x86: checkasm: check for or handle missing cleanup after MMX instructions Not every asm routine is expected clear the MMX state after returning. It is however a requisite for testing floating point code in checkasm. Annotate functions requiring cleanup with declare_func_emms() and issue emms after the call. The remaining functions are checked for having a cleared MMX state after return.
2015-12-11 15:06:38 +02:00
declare_func_emms(AV_CPU_FLAG_MMXEXT, void, uint8_t *src, int topleft, int topright, ptrdiff_t stride);
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
for (pred_mode = 0; pred_mode < 12; pred_mode++) {
if (check_pred_func(h->pred8x8l[pred_mode], "8x8l", pred4x4_modes[codec][pred_mode])) {
int neighbors;
for (neighbors = 0; neighbors <= 0xc000; neighbors += 0x4000) {
int has_topleft = neighbors & 0x8000;
int has_topright = neighbors & 0x4000;
if ((pred_mode == DIAG_DOWN_RIGHT_PRED || pred_mode == VERT_RIGHT_PRED) && !has_topleft)
continue; /* Those aren't allowed according to the spec */
randomize_buffers();
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
call_ref(src0, has_topleft, has_topright, 24*SIZEOF_PIXEL);
call_new(src1, has_topleft, has_topright, 24*SIZEOF_PIXEL);
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
if (memcmp(buf0, buf1, BUF_SIZE))
fail();
checkasm: Explicitly declare function prototypes Now we no longer have to rely on function pointers intentionally declared without specified argument types. This makes it easier to support functions with floating point parameters or return values as well as functions returning 64-bit values on 32-bit architectures. It also avoids having to explicitly cast strides to ptrdiff_t for example.
2015-08-16 13:00:21 +02:00
bench_new(src1, has_topleft, has_topright, 24*SIZEOF_PIXEL);
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
}
}
}
}
}
/* TODO: Add tests for H.264 lossless H/V prediction */
void checkasm_check_h264pred(void)
{
static const struct {
void (*func)(H264PredContext*, uint8_t*, uint8_t*, int, int, int);
const char *name;
} tests[] = {
{ check_pred4x4, "pred4x4" },
{ check_pred8x8, "pred8x8" },
{ check_pred16x16, "pred16x16" },
{ check_pred8x8l, "pred8x8l" },
};
tests/checkasm/h264pred: Use LOCAL_ALIGNED_16() Fixes alignment issue and bus errors Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2015-07-23 00:35:18 +02:00
LOCAL_ALIGNED_16(uint8_t, buf0, [BUF_SIZE]);
LOCAL_ALIGNED_16(uint8_t, buf1, [BUF_SIZE]);
Checkasm: assembly testing and benchmarking tool It provides the following features: * verify correctness by comparing output to the C version. * detect failure to save and restore clobbered callee-saved registers. * detect 32-bit parameters being used as if they were 64-bit in x86-64 (the upper halves are not guaranteed to be zero - but in practice they very often are, which makes those bugs hard to spot otherwise). * easy benchmarking. Compile by running 'make checkasm'. Execute by running 'tests/checkasm/checkasm'. Optional arguments are '--bench' to run benchmarks for all functions, '--bench=<pattern>' to run benchmarks for all functions that starts with <pattern>, and '<integer>' to seed the PRNG for reproducible results. Contains unit tests for most h264pred functions to get started, more tests can be added afterwards using those as a reference. Loosely based on code from x264. Currently only supports x86 and x86-64, but additional architectures shouldn't be too much of an obstacle to add. Note that functions with floating point parameters or floating point return values are not supported. Some compiler-specific features or preprocessor hacks would likely be required to add support for that. Signed-off-by: Janne Grunau <janne-libav@jannau.net>
2015-07-11 20:32:11 +02:00
H264PredContext h;
int test, codec, chroma_format, bit_depth;
for (test = 0; test < FF_ARRAY_ELEMS(tests); test++) {
for (codec = 0; codec < 4; codec++) {
int codec_id = codec_ids[codec];
for (bit_depth = 8; bit_depth <= (codec_id == AV_CODEC_ID_H264 ? 10 : 8); bit_depth++)
for (chroma_format = 1; chroma_format <= (codec_id == AV_CODEC_ID_H264 ? 2 : 1); chroma_format++) {
ff_h264_pred_init(&h, codec_id, bit_depth, chroma_format);
tests[test].func(&h, buf0, buf1, codec, chroma_format, bit_depth);
}
}
report("%s", tests[test].name);
}
}
Reference in New Issue Copy Permalink