generic-library/vpx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
Files
d2697fce4e71a62035e2e6d98e6ea66d2bc4fe48
vpx/test/sad_test.cc
Johann 2967bf355e Merge changes from libvpx/master by cherry-pick 
This commit bring all up-to-date changes from master that are
applicable to nextgenv2. Due to the remove VP10 code in master,
we had to cherry pick the following commits to get those changes:

Add default flags for arm64/armv8 builds

Allows building simple targets with sane default flags.

For example, using the Android arm64 toolchain from the NDK:
https://developer.android.com/ndk/guides/standalone_toolchain.html
./build/tools/make-standalone-toolchain.sh --arch=arm64 \
  --platform=android-24 --install-dir=/tmp/arm64
CROSS=/tmp/arm64/bin/aarch64-linux-android- \
  ~/libvpx/configure --target=arm64-linux-gcc --disable-multithread

BUG=webm:1143

vpx_lpf_horizontal_4_sse2: Remove dead load.

Change-Id: I51026c52baa1f0881fcd5b68e1fdf08a2dc0916e

Fail early when android target does not include --sdk-path

Change-Id: I07e7e63476a2e32e3aae123abdee8b7bbbdc6a8c

configure: clean up var style and set_all usage

Use quotes whenever possible and {} always for variables.

Replace multiple set_all calls with *able_feature().

Conflicts:
	build/make/configure.sh

vp9-svc: Remove some unneeded code/comment.

datarate_test,DatarateTestLarge: normalize bits type

quiets a msvc warning:
conversion from 'const int64_t' to 'size_t', possible loss of data

mips added p6600 cpu support

Removed -funroll-loops

psnr.c: use int64_t for sum of differences

Since the values can be negative.

*.asm: normalize label format

add a trailing ':', though it's optional with the tools we support, it's
more common to use it to mark a label. this also quiets the
orphan-labels warning with nasm/yasm.

BUG=b/29583530

Prevent negative variance

Due to rounding, hbd variance may become negative. This commit put in
check and clamp of negative values to 0.

configure: remove old visual studio support (<2010)

BUG=b/29583530

Conflicts:
	configure

configure: restore vs_version variable

inadvertently lost in the final patchset of:
078dff7 configure: remove old visual studio support (<2010)

this prevents an empty CONFIG_VS_VERSION and avoids make failure

Require x86inc.asm

Force enable x86inc.asm when building for x86. Previously there were
compatibility issues so a flag was added to simplify disabling this
code.

The known issues have been resolved and x86inc.asm is the preferred
abstraction layer (over x86_abi_support.asm).

BUG=b:29583530

convolve_test: fix byte offsets in hbd build

CONVERT_TO_BYTEPTR(x) was corrected in:
003a9d2 Port metric computation changes from nextgenv2
to use the more common (x) within the expansion. offsets should occur
after converting the pointer to the desired type.

+ factorized some common expressions

Conflicts:
	test/convolve_test.cc

vpx_dsp: remove x86inc.asm distinction

BUG=b:29583530

Conflicts:
	vpx_dsp/vpx_dsp.mk
	vpx_dsp/vpx_dsp_rtcd_defs.pl
	vpx_dsp/x86/highbd_variance_sse2.c
	vpx_dsp/x86/variance_sse2.c

test: remove x86inc.asm distinction

BUG=b:29583530

Conflicts:
	test/vp9_subtract_test.cc

configure: remove x86inc.asm distinction

BUG=b:29583530

Change-Id: I59a1192142e89a6a36b906f65a491a734e603617

Update vpx subpixel 1d filter ssse3 asm

Speed test shows the new vertical filters have degradation on Celeron
Chromebook. Added "X86_SUBPIX_VFILTER_PREFER_SLOW_CELERON" to control
the vertical filters activated code. Now just simply active the code
without degradation on Celeron. Later there should be 2 set of vertical
filters ssse3 functions, and let jump table to choose based on CPU type.

improve vpx_filter_block1d* based on replace paddsw+psrlw to pmulhrsw

Make set_reference control API work in VP9

Moved the API patch from NextGenv2. An example was included.
To try it, for example, run the following command:
$ examples/vpx_cx_set_ref vp9 352 288 in.yuv out.ivf 4 30

Conflicts:
	examples.mk
	examples/vpx_cx_set_ref.c
	test/cx_set_ref.sh
	vp9/decoder/vp9_decoder.c

deblock filter : moved from vp8 code branch

The deblocking filters used in vp8 have been moved to vpx_dsp for
use by both vp8 and vp9.

vpx_thread.[hc]: update webp source reference

+ drop the blob hash, the updated reference will be updated in the
commit message

BUG=b/29583578

vpx_thread: use native windows cond var if available

BUG=b/29583578

original webp change:

commit 110ad5835ecd66995d0e7f66dca1b90dea595f5a
Author: James Zern <jzern@google.com>
Date:   Mon Nov 23 19:49:58 2015 -0800

    thread: use native windows cond var if available

    Vista / Server 2008 and up. no speed difference observed.

100644 blob 4fc372b7bc6980a9ed3618c8cce5b67ed7b0f412 src/utils/thread.c
100644 blob 840831185502d42a3246e4b7ff870121c8064791 src/utils/thread.h

vpx_thread: use InitializeCriticalSectionEx if available

BUG=b/29583578

original webp change:

commit 63fadc9ffacc77d4617526a50c696d21d558a70b
Author: James Zern <jzern@google.com>
Date:   Mon Nov 23 20:38:46 2015 -0800

    thread: use InitializeCriticalSectionEx if available

    Windows Vista / Server 2008 and up

100644 blob f84207d89b3a6bb98bfe8f3fa55cad72dfd061ff src/utils/thread.c
100644 blob 840831185502d42a3246e4b7ff870121c8064791 src/utils/thread.h

vpx_thread: use WaitForSingleObjectEx if available

BUG=b/29583578

original webp change:

commit 0fd0e12bfe83f16ce4f1c038b251ccbc13c62ac2
Author: James Zern <jzern@google.com>
Date:   Mon Nov 23 20:40:26 2015 -0800

    thread: use WaitForSingleObjectEx if available

    Windows XP and up

100644 blob d58f74e5523dbc985fc531cf5f0833f1e9157cf0 src/utils/thread.c
100644 blob 840831185502d42a3246e4b7ff870121c8064791 src/utils/thread.h

vpx_thread: use CreateThread for windows phone

BUG=b/29583578

original webp change:

commit d2afe974f9d751de144ef09d31255aea13b442c0
Author: James Zern <jzern@google.com>
Date:   Mon Nov 23 20:41:26 2015 -0800

    thread: use CreateThread for windows phone

    _beginthreadex is unavailable for winrt/uwp

    Change-Id: Ie7412a568278ac67f0047f1764e2521193d74d4d

100644 blob 93f7622797f05f6acc1126e8296c481d276e4047 src/utils/thread.c
100644 blob 840831185502d42a3246e4b7ff870121c8064791 src/utils/thread.h

vp9_postproc.c missing extern.

BUG=webm:1256

deblock: missing const on extern const.

postproc - move filling of noise buffer to vpx_dsp.

Fix encoder crashes for odd size input

clean-up vp9_intrapred_test

remove tuple and overkill VP9IntraPredBase class.

postproc: noise style fixes.

gtest-all.cc: quiet an unused variable warning

under windows / mingw builds

vp9_intrapred_test: follow-up cleanup

address few comments from ce050afaf3

Change-Id: I3eece7efa9335f4210303993ef6c1857ad5c29c8
2016-07-18 10:31:10 -07:00

1025 lines
40 KiB
C++
Raw Blame History

/*
* Copyright (c) 2012 The WebM project authors. All Rights Reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#include <string.h>
#include <limits.h>
#include <stdio.h>
#include "third_party/googletest/src/include/gtest/gtest.h"
#include "./vpx_config.h"
#include "./vpx_dsp_rtcd.h"
#include "test/acm_random.h"
#include "test/clear_system_state.h"
#include "test/register_state_check.h"
#include "test/util.h"
#include "vpx/vpx_codec.h"
#include "vpx_mem/vpx_mem.h"
#include "vpx_ports/mem.h"
typedef unsigned int (*SadMxNFunc)(const uint8_t *src_ptr,
int src_stride,
const uint8_t *ref_ptr,
int ref_stride);
typedef std::tr1::tuple<int, int, SadMxNFunc, int> SadMxNParam;
typedef uint32_t (*SadMxNAvgFunc)(const uint8_t *src_ptr,
int src_stride,
const uint8_t *ref_ptr,
int ref_stride,
const uint8_t *second_pred);
typedef std::tr1::tuple<int, int, SadMxNAvgFunc, int> SadMxNAvgParam;
typedef void (*SadMxNx4Func)(const uint8_t *src_ptr,
int src_stride,
const uint8_t *const ref_ptr[],
int ref_stride,
uint32_t *sad_array);
typedef std::tr1::tuple<int, int, SadMxNx4Func, int> SadMxNx4Param;
using libvpx_test::ACMRandom;
namespace {
class SADTestBase : public ::testing::Test {
public:
SADTestBase(int width, int height, int bit_depth) :
width_(width), height_(height), bd_(bit_depth) {}
static void SetUpTestCase() {
source_data8_ = reinterpret_cast<uint8_t*>(
vpx_memalign(kDataAlignment, kDataBlockSize));
reference_data8_ = reinterpret_cast<uint8_t*>(
vpx_memalign(kDataAlignment, kDataBufferSize));
second_pred8_ = reinterpret_cast<uint8_t*>(
vpx_memalign(kDataAlignment, 128*128));
source_data16_ = reinterpret_cast<uint16_t*>(
vpx_memalign(kDataAlignment, kDataBlockSize*sizeof(uint16_t)));
reference_data16_ = reinterpret_cast<uint16_t*>(
vpx_memalign(kDataAlignment, kDataBufferSize*sizeof(uint16_t)));
second_pred16_ = reinterpret_cast<uint16_t*>(
vpx_memalign(kDataAlignment, 128*128*sizeof(uint16_t)));
}
static void TearDownTestCase() {
vpx_free(source_data8_);
source_data8_ = NULL;
vpx_free(reference_data8_);
reference_data8_ = NULL;
vpx_free(second_pred8_);
second_pred8_ = NULL;
vpx_free(source_data16_);
source_data16_ = NULL;
vpx_free(reference_data16_);
reference_data16_ = NULL;
vpx_free(second_pred16_);
second_pred16_ = NULL;
}
virtual void TearDown() {
libvpx_test::ClearSystemState();
}
protected:
// Handle up to 4 128x128 blocks, with stride up to 256
static const int kDataAlignment = 16;
static const int kDataBlockSize = 128 * 256;
static const int kDataBufferSize = 4 * kDataBlockSize;
virtual void SetUp() {
if (bd_ == -1) {
use_high_bit_depth_ = false;
bit_depth_ = VPX_BITS_8;
source_data_ = source_data8_;
reference_data_ = reference_data8_;
second_pred_ = second_pred8_;
#if CONFIG_VP9_HIGHBITDEPTH
} else {
use_high_bit_depth_ = true;
bit_depth_ = static_cast<vpx_bit_depth_t>(bd_);
source_data_ = CONVERT_TO_BYTEPTR(source_data16_);
reference_data_ = CONVERT_TO_BYTEPTR(reference_data16_);
second_pred_ = CONVERT_TO_BYTEPTR(second_pred16_);
#endif // CONFIG_VP9_HIGHBITDEPTH
}
mask_ = (1 << bit_depth_) - 1;
source_stride_ = (width_ + 31) & ~31;
reference_stride_ = width_ * 2;
rnd_.Reset(ACMRandom::DeterministicSeed());
}
virtual uint8_t *GetReference(int block_idx) {
#if CONFIG_VP9_HIGHBITDEPTH
if (use_high_bit_depth_)
return CONVERT_TO_BYTEPTR(CONVERT_TO_SHORTPTR(reference_data_) +
block_idx * kDataBlockSize);
#endif // CONFIG_VP9_HIGHBITDEPTH
return reference_data_ + block_idx * kDataBlockSize;
}
// Sum of Absolute Differences. Given two blocks, calculate the absolute
// difference between two pixels in the same relative location; accumulate.
unsigned int ReferenceSAD(int block_idx) {
unsigned int sad = 0;
const uint8_t *const reference8 = GetReference(block_idx);
const uint8_t *const source8 = source_data_;
#if CONFIG_VP9_HIGHBITDEPTH
const uint16_t *const reference16 =
CONVERT_TO_SHORTPTR(GetReference(block_idx));
const uint16_t *const source16 = CONVERT_TO_SHORTPTR(source_data_);
#endif // CONFIG_VP9_HIGHBITDEPTH
for (int h = 0; h < height_; ++h) {
for (int w = 0; w < width_; ++w) {
if (!use_high_bit_depth_) {
sad += abs(source8[h * source_stride_ + w] -
reference8[h * reference_stride_ + w]);
#if CONFIG_VP9_HIGHBITDEPTH
} else {
sad += abs(source16[h * source_stride_ + w] -
reference16[h * reference_stride_ + w]);
#endif // CONFIG_VP9_HIGHBITDEPTH
}
}
}
return sad;
}
// Sum of Absolute Differences Average. Given two blocks, and a prediction
// calculate the absolute difference between one pixel and average of the
// corresponding and predicted pixels; accumulate.
unsigned int ReferenceSADavg(int block_idx) {
unsigned int sad = 0;
const uint8_t *const reference8 = GetReference(block_idx);
const uint8_t *const source8 = source_data_;
const uint8_t *const second_pred8 = second_pred_;
#if CONFIG_VP9_HIGHBITDEPTH
const uint16_t *const reference16 =
CONVERT_TO_SHORTPTR(GetReference(block_idx));
const uint16_t *const source16 = CONVERT_TO_SHORTPTR(source_data_);
const uint16_t *const second_pred16 = CONVERT_TO_SHORTPTR(second_pred_);
#endif // CONFIG_VP9_HIGHBITDEPTH
for (int h = 0; h < height_; ++h) {
for (int w = 0; w < width_; ++w) {
if (!use_high_bit_depth_) {
const int tmp = second_pred8[h * width_ + w] +
reference8[h * reference_stride_ + w];
const uint8_t comp_pred = ROUND_POWER_OF_TWO(tmp, 1);
sad += abs(source8[h * source_stride_ + w] - comp_pred);
#if CONFIG_VP9_HIGHBITDEPTH
} else {
const int tmp = second_pred16[h * width_ + w] +
reference16[h * reference_stride_ + w];
const uint16_t comp_pred = ROUND_POWER_OF_TWO(tmp, 1);
sad += abs(source16[h * source_stride_ + w] - comp_pred);
#endif // CONFIG_VP9_HIGHBITDEPTH
}
}
}
return sad;
}
void FillConstant(uint8_t *data, int stride, uint16_t fill_constant) {
uint8_t *data8 = data;
#if CONFIG_VP9_HIGHBITDEPTH
uint16_t *data16 = CONVERT_TO_SHORTPTR(data);
#endif // CONFIG_VP9_HIGHBITDEPTH
for (int h = 0; h < height_; ++h) {
for (int w = 0; w < width_; ++w) {
if (!use_high_bit_depth_) {
data8[h * stride + w] = static_cast<uint8_t>(fill_constant);
#if CONFIG_VP9_HIGHBITDEPTH
} else {
data16[h * stride + w] = fill_constant;
#endif // CONFIG_VP9_HIGHBITDEPTH
}
}
}
}
void FillRandom(uint8_t *data, int stride) {
uint8_t *data8 = data;
#if CONFIG_VP9_HIGHBITDEPTH
uint16_t *data16 = CONVERT_TO_SHORTPTR(data);
#endif // CONFIG_VP9_HIGHBITDEPTH
for (int h = 0; h < height_; ++h) {
for (int w = 0; w < width_; ++w) {
if (!use_high_bit_depth_) {
data8[h * stride + w] = rnd_.Rand8();
#if CONFIG_VP9_HIGHBITDEPTH
} else {
data16[h * stride + w] = rnd_.Rand16() & mask_;
#endif // CONFIG_VP9_HIGHBITDEPTH
}
}
}
}
int width_, height_, mask_, bd_;
vpx_bit_depth_t bit_depth_;
static uint8_t *source_data_;
static uint8_t *reference_data_;
static uint8_t *second_pred_;
int source_stride_;
bool use_high_bit_depth_;
static uint8_t *source_data8_;
static uint8_t *reference_data8_;
static uint8_t *second_pred8_;
static uint16_t *source_data16_;
static uint16_t *reference_data16_;
static uint16_t *second_pred16_;
int reference_stride_;
ACMRandom rnd_;
};
class SADx4Test
: public SADTestBase,
public ::testing::WithParamInterface<SadMxNx4Param> {
public:
SADx4Test() : SADTestBase(GET_PARAM(0), GET_PARAM(1), GET_PARAM(3)) {}
protected:
void SADs(unsigned int *results) {
const uint8_t *references[] = {GetReference(0), GetReference(1),
GetReference(2), GetReference(3)};
ASM_REGISTER_STATE_CHECK(GET_PARAM(2)(source_data_, source_stride_,
references, reference_stride_,
results));
}
void CheckSADs() {
unsigned int reference_sad, exp_sad[4];
SADs(exp_sad);
for (int block = 0; block < 4; ++block) {
reference_sad = ReferenceSAD(block);
EXPECT_EQ(reference_sad, exp_sad[block]) << "block " << block;
}
}
};
class SADTest
: public SADTestBase,
public ::testing::WithParamInterface<SadMxNParam> {
public:
SADTest() : SADTestBase(GET_PARAM(0), GET_PARAM(1), GET_PARAM(3)) {}
protected:
unsigned int SAD(int block_idx) {
unsigned int ret;
const uint8_t *const reference = GetReference(block_idx);
ASM_REGISTER_STATE_CHECK(ret = GET_PARAM(2)(source_data_, source_stride_,
reference, reference_stride_));
return ret;
}
void CheckSAD() {
const unsigned int reference_sad = ReferenceSAD(0);
const unsigned int exp_sad = SAD(0);
ASSERT_EQ(reference_sad, exp_sad);
}
};
class SADavgTest
: public SADTestBase,
public ::testing::WithParamInterface<SadMxNAvgParam> {
public:
SADavgTest() : SADTestBase(GET_PARAM(0), GET_PARAM(1), GET_PARAM(3)) {}
protected:
unsigned int SAD_avg(int block_idx) {
unsigned int ret;
const uint8_t *const reference = GetReference(block_idx);
ASM_REGISTER_STATE_CHECK(ret = GET_PARAM(2)(source_data_, source_stride_,
reference, reference_stride_,
second_pred_));
return ret;
}
void CheckSAD() {
const unsigned int reference_sad = ReferenceSADavg(0);
const unsigned int exp_sad = SAD_avg(0);
ASSERT_EQ(reference_sad, exp_sad);
}
};
uint8_t *SADTestBase::source_data_ = NULL;
uint8_t *SADTestBase::reference_data_ = NULL;
uint8_t *SADTestBase::second_pred_ = NULL;
uint8_t *SADTestBase::source_data8_ = NULL;
uint8_t *SADTestBase::reference_data8_ = NULL;
uint8_t *SADTestBase::second_pred8_ = NULL;
uint16_t *SADTestBase::source_data16_ = NULL;
uint16_t *SADTestBase::reference_data16_ = NULL;
uint16_t *SADTestBase::second_pred16_ = NULL;
TEST_P(SADTest, MaxRef) {
FillConstant(source_data_, source_stride_, 0);
FillConstant(reference_data_, reference_stride_, mask_);
CheckSAD();
}
TEST_P(SADTest, MaxSrc) {
FillConstant(source_data_, source_stride_, mask_);
FillConstant(reference_data_, reference_stride_, 0);
CheckSAD();
}
TEST_P(SADTest, ShortRef) {
const int tmp_stride = reference_stride_;
reference_stride_ >>= 1;
FillRandom(source_data_, source_stride_);
FillRandom(reference_data_, reference_stride_);
CheckSAD();
reference_stride_ = tmp_stride;
}
TEST_P(SADTest, UnalignedRef) {
// The reference frame, but not the source frame, may be unaligned for
// certain types of searches.
const int tmp_stride = reference_stride_;
reference_stride_ -= 1;
FillRandom(source_data_, source_stride_);
FillRandom(reference_data_, reference_stride_);
CheckSAD();
reference_stride_ = tmp_stride;
}
TEST_P(SADTest, ShortSrc) {
const int tmp_stride = source_stride_;
source_stride_ >>= 1;
FillRandom(source_data_, source_stride_);
FillRandom(reference_data_, reference_stride_);
CheckSAD();
source_stride_ = tmp_stride;
}
TEST_P(SADavgTest, MaxRef) {
FillConstant(source_data_, source_stride_, 0);
FillConstant(reference_data_, reference_stride_, mask_);
FillConstant(second_pred_, width_, 0);
CheckSAD();
}
TEST_P(SADavgTest, MaxSrc) {
FillConstant(source_data_, source_stride_, mask_);
FillConstant(reference_data_, reference_stride_, 0);
FillConstant(second_pred_, width_, 0);
CheckSAD();
}
TEST_P(SADavgTest, ShortRef) {
const int tmp_stride = reference_stride_;
reference_stride_ >>= 1;
FillRandom(source_data_, source_stride_);
FillRandom(reference_data_, reference_stride_);
FillRandom(second_pred_, width_);
CheckSAD();
reference_stride_ = tmp_stride;
}
TEST_P(SADavgTest, UnalignedRef) {
// The reference frame, but not the source frame, may be unaligned for
// certain types of searches.
const int tmp_stride = reference_stride_;
reference_stride_ -= 1;
FillRandom(source_data_, source_stride_);
FillRandom(reference_data_, reference_stride_);
FillRandom(second_pred_, width_);
CheckSAD();
reference_stride_ = tmp_stride;
}
TEST_P(SADavgTest, ShortSrc) {
const int tmp_stride = source_stride_;
source_stride_ >>= 1;
FillRandom(source_data_, source_stride_);
FillRandom(reference_data_, reference_stride_);
FillRandom(second_pred_, width_);
CheckSAD();
source_stride_ = tmp_stride;
}
TEST_P(SADx4Test, MaxRef) {
FillConstant(source_data_, source_stride_, 0);
FillConstant(GetReference(0), reference_stride_, mask_);
FillConstant(GetReference(1), reference_stride_, mask_);
FillConstant(GetReference(2), reference_stride_, mask_);
FillConstant(GetReference(3), reference_stride_, mask_);
CheckSADs();
}
TEST_P(SADx4Test, MaxSrc) {
FillConstant(source_data_, source_stride_, mask_);
FillConstant(GetReference(0), reference_stride_, 0);
FillConstant(GetReference(1), reference_stride_, 0);
FillConstant(GetReference(2), reference_stride_, 0);
FillConstant(GetReference(3), reference_stride_, 0);
CheckSADs();
}
TEST_P(SADx4Test, ShortRef) {
int tmp_stride = reference_stride_;
reference_stride_ >>= 1;
FillRandom(source_data_, source_stride_);
FillRandom(GetReference(0), reference_stride_);
FillRandom(GetReference(1), reference_stride_);
FillRandom(GetReference(2), reference_stride_);
FillRandom(GetReference(3), reference_stride_);
CheckSADs();
reference_stride_ = tmp_stride;
}
TEST_P(SADx4Test, UnalignedRef) {
// The reference frame, but not the source frame, may be unaligned for
// certain types of searches.
int tmp_stride = reference_stride_;
reference_stride_ -= 1;
FillRandom(source_data_, source_stride_);
FillRandom(GetReference(0), reference_stride_);
FillRandom(GetReference(1), reference_stride_);
FillRandom(GetReference(2), reference_stride_);
FillRandom(GetReference(3), reference_stride_);
CheckSADs();
reference_stride_ = tmp_stride;
}
TEST_P(SADx4Test, ShortSrc) {
int tmp_stride = source_stride_;
source_stride_ >>= 1;
FillRandom(source_data_, source_stride_);
FillRandom(GetReference(0), reference_stride_);
FillRandom(GetReference(1), reference_stride_);
FillRandom(GetReference(2), reference_stride_);
FillRandom(GetReference(3), reference_stride_);
CheckSADs();
source_stride_ = tmp_stride;
}
TEST_P(SADx4Test, SrcAlignedByWidth) {
uint8_t * tmp_source_data = source_data_;
source_data_ += width_;
FillRandom(source_data_, source_stride_);
FillRandom(GetReference(0), reference_stride_);
FillRandom(GetReference(1), reference_stride_);
FillRandom(GetReference(2), reference_stride_);
FillRandom(GetReference(3), reference_stride_);
CheckSADs();
source_data_ = tmp_source_data;
}
using std::tr1::make_tuple;
//------------------------------------------------------------------------------
// C functions
const SadMxNParam c_tests[] = {
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_sad128x128_c, -1),
make_tuple(128, 64, &vpx_sad128x64_c, -1),
make_tuple(64, 128, &vpx_sad64x128_c, -1),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_sad64x64_c, -1),
make_tuple(64, 32, &vpx_sad64x32_c, -1),
make_tuple(32, 64, &vpx_sad32x64_c, -1),
make_tuple(32, 32, &vpx_sad32x32_c, -1),
make_tuple(32, 16, &vpx_sad32x16_c, -1),
make_tuple(16, 32, &vpx_sad16x32_c, -1),
make_tuple(16, 16, &vpx_sad16x16_c, -1),
make_tuple(16, 8, &vpx_sad16x8_c, -1),
make_tuple(8, 16, &vpx_sad8x16_c, -1),
make_tuple(8, 8, &vpx_sad8x8_c, -1),
make_tuple(8, 4, &vpx_sad8x4_c, -1),
make_tuple(4, 8, &vpx_sad4x8_c, -1),
make_tuple(4, 4, &vpx_sad4x4_c, -1),
#if CONFIG_VP9_HIGHBITDEPTH
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_highbd_sad128x128_c, 8),
make_tuple(128, 64, &vpx_highbd_sad128x64_c, 8),
make_tuple(64, 128, &vpx_highbd_sad64x128_c, 8),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_highbd_sad64x64_c, 8),
make_tuple(64, 32, &vpx_highbd_sad64x32_c, 8),
make_tuple(32, 64, &vpx_highbd_sad32x64_c, 8),
make_tuple(32, 32, &vpx_highbd_sad32x32_c, 8),
make_tuple(32, 16, &vpx_highbd_sad32x16_c, 8),
make_tuple(16, 32, &vpx_highbd_sad16x32_c, 8),
make_tuple(16, 16, &vpx_highbd_sad16x16_c, 8),
make_tuple(16, 8, &vpx_highbd_sad16x8_c, 8),
make_tuple(8, 16, &vpx_highbd_sad8x16_c, 8),
make_tuple(8, 8, &vpx_highbd_sad8x8_c, 8),
make_tuple(8, 4, &vpx_highbd_sad8x4_c, 8),
make_tuple(4, 8, &vpx_highbd_sad4x8_c, 8),
make_tuple(4, 4, &vpx_highbd_sad4x4_c, 8),
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_highbd_sad128x128_c, 10),
make_tuple(128, 64, &vpx_highbd_sad128x64_c, 10),
make_tuple(64, 128, &vpx_highbd_sad64x128_c, 10),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_highbd_sad64x64_c, 10),
make_tuple(64, 32, &vpx_highbd_sad64x32_c, 10),
make_tuple(32, 64, &vpx_highbd_sad32x64_c, 10),
make_tuple(32, 32, &vpx_highbd_sad32x32_c, 10),
make_tuple(32, 16, &vpx_highbd_sad32x16_c, 10),
make_tuple(16, 32, &vpx_highbd_sad16x32_c, 10),
make_tuple(16, 16, &vpx_highbd_sad16x16_c, 10),
make_tuple(16, 8, &vpx_highbd_sad16x8_c, 10),
make_tuple(8, 16, &vpx_highbd_sad8x16_c, 10),
make_tuple(8, 8, &vpx_highbd_sad8x8_c, 10),
make_tuple(8, 4, &vpx_highbd_sad8x4_c, 10),
make_tuple(4, 8, &vpx_highbd_sad4x8_c, 10),
make_tuple(4, 4, &vpx_highbd_sad4x4_c, 10),
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_highbd_sad128x128_c, 12),
make_tuple(128, 64, &vpx_highbd_sad128x64_c, 12),
make_tuple(64, 128, &vpx_highbd_sad64x128_c, 12),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_highbd_sad64x64_c, 12),
make_tuple(64, 32, &vpx_highbd_sad64x32_c, 12),
make_tuple(32, 64, &vpx_highbd_sad32x64_c, 12),
make_tuple(32, 32, &vpx_highbd_sad32x32_c, 12),
make_tuple(32, 16, &vpx_highbd_sad32x16_c, 12),
make_tuple(16, 32, &vpx_highbd_sad16x32_c, 12),
make_tuple(16, 16, &vpx_highbd_sad16x16_c, 12),
make_tuple(16, 8, &vpx_highbd_sad16x8_c, 12),
make_tuple(8, 16, &vpx_highbd_sad8x16_c, 12),
make_tuple(8, 8, &vpx_highbd_sad8x8_c, 12),
make_tuple(8, 4, &vpx_highbd_sad8x4_c, 12),
make_tuple(4, 8, &vpx_highbd_sad4x8_c, 12),
make_tuple(4, 4, &vpx_highbd_sad4x4_c, 12),
#endif // CONFIG_VP9_HIGHBITDEPTH
};
INSTANTIATE_TEST_CASE_P(C, SADTest, ::testing::ValuesIn(c_tests));
const SadMxNAvgParam avg_c_tests[] = {
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_sad128x128_avg_c, -1),
make_tuple(128, 64, &vpx_sad128x64_avg_c, -1),
make_tuple(64, 128, &vpx_sad64x128_avg_c, -1),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_sad64x64_avg_c, -1),
make_tuple(64, 32, &vpx_sad64x32_avg_c, -1),
make_tuple(32, 64, &vpx_sad32x64_avg_c, -1),
make_tuple(32, 32, &vpx_sad32x32_avg_c, -1),
make_tuple(32, 16, &vpx_sad32x16_avg_c, -1),
make_tuple(16, 32, &vpx_sad16x32_avg_c, -1),
make_tuple(16, 16, &vpx_sad16x16_avg_c, -1),
make_tuple(16, 8, &vpx_sad16x8_avg_c, -1),
make_tuple(8, 16, &vpx_sad8x16_avg_c, -1),
make_tuple(8, 8, &vpx_sad8x8_avg_c, -1),
make_tuple(8, 4, &vpx_sad8x4_avg_c, -1),
make_tuple(4, 8, &vpx_sad4x8_avg_c, -1),
make_tuple(4, 4, &vpx_sad4x4_avg_c, -1),
#if CONFIG_VP9_HIGHBITDEPTH
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_highbd_sad128x128_avg_c, 8),
make_tuple(128, 64, &vpx_highbd_sad128x64_avg_c, 8),
make_tuple(64, 128, &vpx_highbd_sad64x128_avg_c, 8),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_highbd_sad64x64_avg_c, 8),
make_tuple(64, 32, &vpx_highbd_sad64x32_avg_c, 8),
make_tuple(32, 64, &vpx_highbd_sad32x64_avg_c, 8),
make_tuple(32, 32, &vpx_highbd_sad32x32_avg_c, 8),
make_tuple(32, 16, &vpx_highbd_sad32x16_avg_c, 8),
make_tuple(16, 32, &vpx_highbd_sad16x32_avg_c, 8),
make_tuple(16, 16, &vpx_highbd_sad16x16_avg_c, 8),
make_tuple(16, 8, &vpx_highbd_sad16x8_avg_c, 8),
make_tuple(8, 16, &vpx_highbd_sad8x16_avg_c, 8),
make_tuple(8, 8, &vpx_highbd_sad8x8_avg_c, 8),
make_tuple(8, 4, &vpx_highbd_sad8x4_avg_c, 8),
make_tuple(4, 8, &vpx_highbd_sad4x8_avg_c, 8),
make_tuple(4, 4, &vpx_highbd_sad4x4_avg_c, 8),
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_highbd_sad128x128_avg_c, 10),
make_tuple(128, 64, &vpx_highbd_sad128x64_avg_c, 10),
make_tuple(64, 128, &vpx_highbd_sad64x128_avg_c, 10),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_highbd_sad64x64_avg_c, 10),
make_tuple(64, 32, &vpx_highbd_sad64x32_avg_c, 10),
make_tuple(32, 64, &vpx_highbd_sad32x64_avg_c, 10),
make_tuple(32, 32, &vpx_highbd_sad32x32_avg_c, 10),
make_tuple(32, 16, &vpx_highbd_sad32x16_avg_c, 10),
make_tuple(16, 32, &vpx_highbd_sad16x32_avg_c, 10),
make_tuple(16, 16, &vpx_highbd_sad16x16_avg_c, 10),
make_tuple(16, 8, &vpx_highbd_sad16x8_avg_c, 10),
make_tuple(8, 16, &vpx_highbd_sad8x16_avg_c, 10),
make_tuple(8, 8, &vpx_highbd_sad8x8_avg_c, 10),
make_tuple(8, 4, &vpx_highbd_sad8x4_avg_c, 10),
make_tuple(4, 8, &vpx_highbd_sad4x8_avg_c, 10),
make_tuple(4, 4, &vpx_highbd_sad4x4_avg_c, 10),
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_highbd_sad128x128_avg_c, 12),
make_tuple(128, 64, &vpx_highbd_sad128x64_avg_c, 12),
make_tuple(64, 128, &vpx_highbd_sad64x128_avg_c, 12),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_highbd_sad64x64_avg_c, 12),
make_tuple(64, 32, &vpx_highbd_sad64x32_avg_c, 12),
make_tuple(32, 64, &vpx_highbd_sad32x64_avg_c, 12),
make_tuple(32, 32, &vpx_highbd_sad32x32_avg_c, 12),
make_tuple(32, 16, &vpx_highbd_sad32x16_avg_c, 12),
make_tuple(16, 32, &vpx_highbd_sad16x32_avg_c, 12),
make_tuple(16, 16, &vpx_highbd_sad16x16_avg_c, 12),
make_tuple(16, 8, &vpx_highbd_sad16x8_avg_c, 12),
make_tuple(8, 16, &vpx_highbd_sad8x16_avg_c, 12),
make_tuple(8, 8, &vpx_highbd_sad8x8_avg_c, 12),
make_tuple(8, 4, &vpx_highbd_sad8x4_avg_c, 12),
make_tuple(4, 8, &vpx_highbd_sad4x8_avg_c, 12),
make_tuple(4, 4, &vpx_highbd_sad4x4_avg_c, 12),
#endif // CONFIG_VP9_HIGHBITDEPTH
};
INSTANTIATE_TEST_CASE_P(C, SADavgTest, ::testing::ValuesIn(avg_c_tests));
const SadMxNx4Param x4d_c_tests[] = {
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_sad128x128x4d_c, -1),
make_tuple(128, 64, &vpx_sad128x64x4d_c, -1),
make_tuple(64, 128, &vpx_sad64x128x4d_c, -1),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_sad64x64x4d_c, -1),
make_tuple(64, 32, &vpx_sad64x32x4d_c, -1),
make_tuple(32, 64, &vpx_sad32x64x4d_c, -1),
make_tuple(32, 32, &vpx_sad32x32x4d_c, -1),
make_tuple(32, 16, &vpx_sad32x16x4d_c, -1),
make_tuple(16, 32, &vpx_sad16x32x4d_c, -1),
make_tuple(16, 16, &vpx_sad16x16x4d_c, -1),
make_tuple(16, 8, &vpx_sad16x8x4d_c, -1),
make_tuple(8, 16, &vpx_sad8x16x4d_c, -1),
make_tuple(8, 8, &vpx_sad8x8x4d_c, -1),
make_tuple(8, 4, &vpx_sad8x4x4d_c, -1),
make_tuple(4, 8, &vpx_sad4x8x4d_c, -1),
make_tuple(4, 4, &vpx_sad4x4x4d_c, -1),
#if CONFIG_VP9_HIGHBITDEPTH
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_highbd_sad128x128x4d_c, 8),
make_tuple(128, 64, &vpx_highbd_sad128x64x4d_c, 8),
make_tuple(64, 128, &vpx_highbd_sad64x128x4d_c, 8),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_highbd_sad64x64x4d_c, 8),
make_tuple(64, 32, &vpx_highbd_sad64x32x4d_c, 8),
make_tuple(32, 64, &vpx_highbd_sad32x64x4d_c, 8),
make_tuple(32, 32, &vpx_highbd_sad32x32x4d_c, 8),
make_tuple(32, 16, &vpx_highbd_sad32x16x4d_c, 8),
make_tuple(16, 32, &vpx_highbd_sad16x32x4d_c, 8),
make_tuple(16, 16, &vpx_highbd_sad16x16x4d_c, 8),
make_tuple(16, 8, &vpx_highbd_sad16x8x4d_c, 8),
make_tuple(8, 16, &vpx_highbd_sad8x16x4d_c, 8),
make_tuple(8, 8, &vpx_highbd_sad8x8x4d_c, 8),
make_tuple(8, 4, &vpx_highbd_sad8x4x4d_c, 8),
make_tuple(4, 8, &vpx_highbd_sad4x8x4d_c, 8),
make_tuple(4, 4, &vpx_highbd_sad4x4x4d_c, 8),
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_highbd_sad128x128x4d_c, 10),
make_tuple(128, 64, &vpx_highbd_sad128x64x4d_c, 10),
make_tuple(64, 128, &vpx_highbd_sad64x128x4d_c, 10),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_highbd_sad64x64x4d_c, 10),
make_tuple(64, 32, &vpx_highbd_sad64x32x4d_c, 10),
make_tuple(32, 64, &vpx_highbd_sad32x64x4d_c, 10),
make_tuple(32, 32, &vpx_highbd_sad32x32x4d_c, 10),
make_tuple(32, 16, &vpx_highbd_sad32x16x4d_c, 10),
make_tuple(16, 32, &vpx_highbd_sad16x32x4d_c, 10),
make_tuple(16, 16, &vpx_highbd_sad16x16x4d_c, 10),
make_tuple(16, 8, &vpx_highbd_sad16x8x4d_c, 10),
make_tuple(8, 16, &vpx_highbd_sad8x16x4d_c, 10),
make_tuple(8, 8, &vpx_highbd_sad8x8x4d_c, 10),
make_tuple(8, 4, &vpx_highbd_sad8x4x4d_c, 10),
make_tuple(4, 8, &vpx_highbd_sad4x8x4d_c, 10),
make_tuple(4, 4, &vpx_highbd_sad4x4x4d_c, 10),
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_highbd_sad128x128x4d_c, 12),
make_tuple(128, 64, &vpx_highbd_sad128x64x4d_c, 12),
make_tuple(64, 128, &vpx_highbd_sad64x128x4d_c, 12),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_highbd_sad64x64x4d_c, 12),
make_tuple(64, 32, &vpx_highbd_sad64x32x4d_c, 12),
make_tuple(32, 64, &vpx_highbd_sad32x64x4d_c, 12),
make_tuple(32, 32, &vpx_highbd_sad32x32x4d_c, 12),
make_tuple(32, 16, &vpx_highbd_sad32x16x4d_c, 12),
make_tuple(16, 32, &vpx_highbd_sad16x32x4d_c, 12),
make_tuple(16, 16, &vpx_highbd_sad16x16x4d_c, 12),
make_tuple(16, 8, &vpx_highbd_sad16x8x4d_c, 12),
make_tuple(8, 16, &vpx_highbd_sad8x16x4d_c, 12),
make_tuple(8, 8, &vpx_highbd_sad8x8x4d_c, 12),
make_tuple(8, 4, &vpx_highbd_sad8x4x4d_c, 12),
make_tuple(4, 8, &vpx_highbd_sad4x8x4d_c, 12),
make_tuple(4, 4, &vpx_highbd_sad4x4x4d_c, 12),
#endif // CONFIG_VP9_HIGHBITDEPTH
};
INSTANTIATE_TEST_CASE_P(C, SADx4Test, ::testing::ValuesIn(x4d_c_tests));
//------------------------------------------------------------------------------
// ARM functions
#if HAVE_MEDIA
const SadMxNParam media_tests[] = {
make_tuple(16, 16, &vpx_sad16x16_media, -1),
};
INSTANTIATE_TEST_CASE_P(MEDIA, SADTest, ::testing::ValuesIn(media_tests));
#endif // HAVE_MEDIA
#if HAVE_NEON
const SadMxNParam neon_tests[] = {
make_tuple(64, 64, &vpx_sad64x64_neon, -1),
make_tuple(32, 32, &vpx_sad32x32_neon, -1),
make_tuple(16, 16, &vpx_sad16x16_neon, -1),
make_tuple(16, 8, &vpx_sad16x8_neon, -1),
make_tuple(8, 16, &vpx_sad8x16_neon, -1),
make_tuple(8, 8, &vpx_sad8x8_neon, -1),
make_tuple(4, 4, &vpx_sad4x4_neon, -1),
};
INSTANTIATE_TEST_CASE_P(NEON, SADTest, ::testing::ValuesIn(neon_tests));
const SadMxNx4Param x4d_neon_tests[] = {
make_tuple(64, 64, &vpx_sad64x64x4d_neon, -1),
make_tuple(32, 32, &vpx_sad32x32x4d_neon, -1),
make_tuple(16, 16, &vpx_sad16x16x4d_neon, -1),
};
INSTANTIATE_TEST_CASE_P(NEON, SADx4Test, ::testing::ValuesIn(x4d_neon_tests));
#endif // HAVE_NEON
//------------------------------------------------------------------------------
// x86 functions
#if HAVE_SSE2
const SadMxNParam sse2_tests[] = {
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_sad128x128_sse2, -1),
make_tuple(128, 64, &vpx_sad128x64_sse2, -1),
make_tuple(64, 128, &vpx_sad64x128_sse2, -1),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_sad64x64_sse2, -1),
make_tuple(64, 32, &vpx_sad64x32_sse2, -1),
make_tuple(32, 64, &vpx_sad32x64_sse2, -1),
make_tuple(32, 32, &vpx_sad32x32_sse2, -1),
make_tuple(32, 16, &vpx_sad32x16_sse2, -1),
make_tuple(16, 32, &vpx_sad16x32_sse2, -1),
make_tuple(16, 16, &vpx_sad16x16_sse2, -1),
make_tuple(16, 8, &vpx_sad16x8_sse2, -1),
make_tuple(8, 16, &vpx_sad8x16_sse2, -1),
make_tuple(8, 8, &vpx_sad8x8_sse2, -1),
make_tuple(8, 4, &vpx_sad8x4_sse2, -1),
make_tuple(4, 8, &vpx_sad4x8_sse2, -1),
make_tuple(4, 4, &vpx_sad4x4_sse2, -1),
#if CONFIG_VP9_HIGHBITDEPTH
make_tuple(64, 64, &vpx_highbd_sad64x64_sse2, 8),
make_tuple(64, 32, &vpx_highbd_sad64x32_sse2, 8),
make_tuple(32, 64, &vpx_highbd_sad32x64_sse2, 8),
make_tuple(32, 32, &vpx_highbd_sad32x32_sse2, 8),
make_tuple(32, 16, &vpx_highbd_sad32x16_sse2, 8),
make_tuple(16, 32, &vpx_highbd_sad16x32_sse2, 8),
make_tuple(16, 16, &vpx_highbd_sad16x16_sse2, 8),
make_tuple(16, 8, &vpx_highbd_sad16x8_sse2, 8),
make_tuple(8, 16, &vpx_highbd_sad8x16_sse2, 8),
make_tuple(8, 8, &vpx_highbd_sad8x8_sse2, 8),
make_tuple(8, 4, &vpx_highbd_sad8x4_sse2, 8),
make_tuple(64, 64, &vpx_highbd_sad64x64_sse2, 10),
make_tuple(64, 32, &vpx_highbd_sad64x32_sse2, 10),
make_tuple(32, 64, &vpx_highbd_sad32x64_sse2, 10),
make_tuple(32, 32, &vpx_highbd_sad32x32_sse2, 10),
make_tuple(32, 16, &vpx_highbd_sad32x16_sse2, 10),
make_tuple(16, 32, &vpx_highbd_sad16x32_sse2, 10),
make_tuple(16, 16, &vpx_highbd_sad16x16_sse2, 10),
make_tuple(16, 8, &vpx_highbd_sad16x8_sse2, 10),
make_tuple(8, 16, &vpx_highbd_sad8x16_sse2, 10),
make_tuple(8, 8, &vpx_highbd_sad8x8_sse2, 10),
make_tuple(8, 4, &vpx_highbd_sad8x4_sse2, 10),
make_tuple(64, 64, &vpx_highbd_sad64x64_sse2, 12),
make_tuple(64, 32, &vpx_highbd_sad64x32_sse2, 12),
make_tuple(32, 64, &vpx_highbd_sad32x64_sse2, 12),
make_tuple(32, 32, &vpx_highbd_sad32x32_sse2, 12),
make_tuple(32, 16, &vpx_highbd_sad32x16_sse2, 12),
make_tuple(16, 32, &vpx_highbd_sad16x32_sse2, 12),
make_tuple(16, 16, &vpx_highbd_sad16x16_sse2, 12),
make_tuple(16, 8, &vpx_highbd_sad16x8_sse2, 12),
make_tuple(8, 16, &vpx_highbd_sad8x16_sse2, 12),
make_tuple(8, 8, &vpx_highbd_sad8x8_sse2, 12),
make_tuple(8, 4, &vpx_highbd_sad8x4_sse2, 12),
#endif // CONFIG_VP9_HIGHBITDEPTH
};
INSTANTIATE_TEST_CASE_P(SSE2, SADTest, ::testing::ValuesIn(sse2_tests));
const SadMxNAvgParam avg_sse2_tests[] = {
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_sad128x128_avg_sse2, -1),
make_tuple(128, 64, &vpx_sad128x64_avg_sse2, -1),
make_tuple(64, 128, &vpx_sad64x128_avg_sse2, -1),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_sad64x64_avg_sse2, -1),
make_tuple(64, 32, &vpx_sad64x32_avg_sse2, -1),
make_tuple(32, 64, &vpx_sad32x64_avg_sse2, -1),
make_tuple(32, 32, &vpx_sad32x32_avg_sse2, -1),
make_tuple(32, 16, &vpx_sad32x16_avg_sse2, -1),
make_tuple(16, 32, &vpx_sad16x32_avg_sse2, -1),
make_tuple(16, 16, &vpx_sad16x16_avg_sse2, -1),
make_tuple(16, 8, &vpx_sad16x8_avg_sse2, -1),
make_tuple(8, 16, &vpx_sad8x16_avg_sse2, -1),
make_tuple(8, 8, &vpx_sad8x8_avg_sse2, -1),
make_tuple(8, 4, &vpx_sad8x4_avg_sse2, -1),
make_tuple(4, 8, &vpx_sad4x8_avg_sse2, -1),
make_tuple(4, 4, &vpx_sad4x4_avg_sse2, -1),
#if CONFIG_VP9_HIGHBITDEPTH
make_tuple(64, 64, &vpx_highbd_sad64x64_avg_sse2, 8),
make_tuple(64, 32, &vpx_highbd_sad64x32_avg_sse2, 8),
make_tuple(32, 64, &vpx_highbd_sad32x64_avg_sse2, 8),
make_tuple(32, 32, &vpx_highbd_sad32x32_avg_sse2, 8),
make_tuple(32, 16, &vpx_highbd_sad32x16_avg_sse2, 8),
make_tuple(16, 32, &vpx_highbd_sad16x32_avg_sse2, 8),
make_tuple(16, 16, &vpx_highbd_sad16x16_avg_sse2, 8),
make_tuple(16, 8, &vpx_highbd_sad16x8_avg_sse2, 8),
make_tuple(8, 16, &vpx_highbd_sad8x16_avg_sse2, 8),
make_tuple(8, 8, &vpx_highbd_sad8x8_avg_sse2, 8),
make_tuple(8, 4, &vpx_highbd_sad8x4_avg_sse2, 8),
make_tuple(64, 64, &vpx_highbd_sad64x64_avg_sse2, 10),
make_tuple(64, 32, &vpx_highbd_sad64x32_avg_sse2, 10),
make_tuple(32, 64, &vpx_highbd_sad32x64_avg_sse2, 10),
make_tuple(32, 32, &vpx_highbd_sad32x32_avg_sse2, 10),
make_tuple(32, 16, &vpx_highbd_sad32x16_avg_sse2, 10),
make_tuple(16, 32, &vpx_highbd_sad16x32_avg_sse2, 10),
make_tuple(16, 16, &vpx_highbd_sad16x16_avg_sse2, 10),
make_tuple(16, 8, &vpx_highbd_sad16x8_avg_sse2, 10),
make_tuple(8, 16, &vpx_highbd_sad8x16_avg_sse2, 10),
make_tuple(8, 8, &vpx_highbd_sad8x8_avg_sse2, 10),
make_tuple(8, 4, &vpx_highbd_sad8x4_avg_sse2, 10),
make_tuple(64, 64, &vpx_highbd_sad64x64_avg_sse2, 12),
make_tuple(64, 32, &vpx_highbd_sad64x32_avg_sse2, 12),
make_tuple(32, 64, &vpx_highbd_sad32x64_avg_sse2, 12),
make_tuple(32, 32, &vpx_highbd_sad32x32_avg_sse2, 12),
make_tuple(32, 16, &vpx_highbd_sad32x16_avg_sse2, 12),
make_tuple(16, 32, &vpx_highbd_sad16x32_avg_sse2, 12),
make_tuple(16, 16, &vpx_highbd_sad16x16_avg_sse2, 12),
make_tuple(16, 8, &vpx_highbd_sad16x8_avg_sse2, 12),
make_tuple(8, 16, &vpx_highbd_sad8x16_avg_sse2, 12),
make_tuple(8, 8, &vpx_highbd_sad8x8_avg_sse2, 12),
make_tuple(8, 4, &vpx_highbd_sad8x4_avg_sse2, 12),
#endif // CONFIG_VP9_HIGHBITDEPTH
};
INSTANTIATE_TEST_CASE_P(SSE2, SADavgTest, ::testing::ValuesIn(avg_sse2_tests));
const SadMxNx4Param x4d_sse2_tests[] = {
#if CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(128, 128, &vpx_sad128x128x4d_sse2, -1),
make_tuple(128, 64, &vpx_sad128x64x4d_sse2, -1),
make_tuple(64, 128, &vpx_sad64x128x4d_sse2, -1),
#endif // CONFIG_VP10 && CONFIG_EXT_PARTITION
make_tuple(64, 64, &vpx_sad64x64x4d_sse2, -1),
make_tuple(64, 32, &vpx_sad64x32x4d_sse2, -1),
make_tuple(32, 64, &vpx_sad32x64x4d_sse2, -1),
make_tuple(32, 32, &vpx_sad32x32x4d_sse2, -1),
make_tuple(32, 16, &vpx_sad32x16x4d_sse2, -1),
make_tuple(16, 32, &vpx_sad16x32x4d_sse2, -1),
make_tuple(16, 16, &vpx_sad16x16x4d_sse2, -1),
make_tuple(16, 8, &vpx_sad16x8x4d_sse2, -1),
make_tuple(8, 16, &vpx_sad8x16x4d_sse2, -1),
make_tuple(8, 8, &vpx_sad8x8x4d_sse2, -1),
make_tuple(8, 4, &vpx_sad8x4x4d_sse2, -1),
make_tuple(4, 8, &vpx_sad4x8x4d_sse2, -1),
make_tuple(4, 4, &vpx_sad4x4x4d_sse2, -1),
#if CONFIG_VP9_HIGHBITDEPTH
make_tuple(64, 64, &vpx_highbd_sad64x64x4d_sse2, 8),
make_tuple(64, 32, &vpx_highbd_sad64x32x4d_sse2, 8),
make_tuple(32, 64, &vpx_highbd_sad32x64x4d_sse2, 8),
make_tuple(32, 32, &vpx_highbd_sad32x32x4d_sse2, 8),
make_tuple(32, 16, &vpx_highbd_sad32x16x4d_sse2, 8),
make_tuple(16, 32, &vpx_highbd_sad16x32x4d_sse2, 8),
make_tuple(16, 16, &vpx_highbd_sad16x16x4d_sse2, 8),
make_tuple(16, 8, &vpx_highbd_sad16x8x4d_sse2, 8),
make_tuple(8, 16, &vpx_highbd_sad8x16x4d_sse2, 8),
make_tuple(8, 8, &vpx_highbd_sad8x8x4d_sse2, 8),
make_tuple(8, 4, &vpx_highbd_sad8x4x4d_sse2, 8),
make_tuple(4, 8, &vpx_highbd_sad4x8x4d_sse2, 8),
make_tuple(4, 4, &vpx_highbd_sad4x4x4d_sse2, 8),
make_tuple(64, 64, &vpx_highbd_sad64x64x4d_sse2, 10),
make_tuple(64, 32, &vpx_highbd_sad64x32x4d_sse2, 10),
make_tuple(32, 64, &vpx_highbd_sad32x64x4d_sse2, 10),
make_tuple(32, 32, &vpx_highbd_sad32x32x4d_sse2, 10),
make_tuple(32, 16, &vpx_highbd_sad32x16x4d_sse2, 10),
make_tuple(16, 32, &vpx_highbd_sad16x32x4d_sse2, 10),
make_tuple(16, 16, &vpx_highbd_sad16x16x4d_sse2, 10),
make_tuple(16, 8, &vpx_highbd_sad16x8x4d_sse2, 10),
make_tuple(8, 16, &vpx_highbd_sad8x16x4d_sse2, 10),
make_tuple(8, 8, &vpx_highbd_sad8x8x4d_sse2, 10),
make_tuple(8, 4, &vpx_highbd_sad8x4x4d_sse2, 10),
make_tuple(4, 8, &vpx_highbd_sad4x8x4d_sse2, 10),
make_tuple(4, 4, &vpx_highbd_sad4x4x4d_sse2, 10),
make_tuple(64, 64, &vpx_highbd_sad64x64x4d_sse2, 12),
make_tuple(64, 32, &vpx_highbd_sad64x32x4d_sse2, 12),
make_tuple(32, 64, &vpx_highbd_sad32x64x4d_sse2, 12),
make_tuple(32, 32, &vpx_highbd_sad32x32x4d_sse2, 12),
make_tuple(32, 16, &vpx_highbd_sad32x16x4d_sse2, 12),
make_tuple(16, 32, &vpx_highbd_sad16x32x4d_sse2, 12),
make_tuple(16, 16, &vpx_highbd_sad16x16x4d_sse2, 12),
make_tuple(16, 8, &vpx_highbd_sad16x8x4d_sse2, 12),
make_tuple(8, 16, &vpx_highbd_sad8x16x4d_sse2, 12),
make_tuple(8, 8, &vpx_highbd_sad8x8x4d_sse2, 12),
make_tuple(8, 4, &vpx_highbd_sad8x4x4d_sse2, 12),
make_tuple(4, 8, &vpx_highbd_sad4x8x4d_sse2, 12),
make_tuple(4, 4, &vpx_highbd_sad4x4x4d_sse2, 12),
#endif // CONFIG_VP9_HIGHBITDEPTH
};
INSTANTIATE_TEST_CASE_P(SSE2, SADx4Test, ::testing::ValuesIn(x4d_sse2_tests));
#endif // HAVE_SSE2
#if HAVE_SSE3
// Only functions are x3, which do not have tests.
#endif // HAVE_SSE3
#if HAVE_SSSE3
// Only functions are x3, which do not have tests.
#endif // HAVE_SSSE3
#if HAVE_SSE4_1
// Only functions are x8, which do not have tests.
#endif // HAVE_SSE4_1
#if HAVE_AVX2
const SadMxNParam avx2_tests[] = {
make_tuple(64, 64, &vpx_sad64x64_avx2, -1),
make_tuple(64, 32, &vpx_sad64x32_avx2, -1),
make_tuple(32, 64, &vpx_sad32x64_avx2, -1),
make_tuple(32, 32, &vpx_sad32x32_avx2, -1),
make_tuple(32, 16, &vpx_sad32x16_avx2, -1),
};
INSTANTIATE_TEST_CASE_P(AVX2, SADTest, ::testing::ValuesIn(avx2_tests));
const SadMxNAvgParam avg_avx2_tests[] = {
make_tuple(64, 64, &vpx_sad64x64_avg_avx2, -1),
make_tuple(64, 32, &vpx_sad64x32_avg_avx2, -1),
make_tuple(32, 64, &vpx_sad32x64_avg_avx2, -1),
make_tuple(32, 32, &vpx_sad32x32_avg_avx2, -1),
make_tuple(32, 16, &vpx_sad32x16_avg_avx2, -1),
};
INSTANTIATE_TEST_CASE_P(AVX2, SADavgTest, ::testing::ValuesIn(avg_avx2_tests));
const SadMxNx4Param x4d_avx2_tests[] = {
make_tuple(64, 64, &vpx_sad64x64x4d_avx2, -1),
make_tuple(32, 32, &vpx_sad32x32x4d_avx2, -1),
};
INSTANTIATE_TEST_CASE_P(AVX2, SADx4Test, ::testing::ValuesIn(x4d_avx2_tests));
#endif // HAVE_AVX2
//------------------------------------------------------------------------------
// MIPS functions
#if HAVE_MSA
const SadMxNParam msa_tests[] = {
make_tuple(64, 64, &vpx_sad64x64_msa, -1),
make_tuple(64, 32, &vpx_sad64x32_msa, -1),
make_tuple(32, 64, &vpx_sad32x64_msa, -1),
make_tuple(32, 32, &vpx_sad32x32_msa, -1),
make_tuple(32, 16, &vpx_sad32x16_msa, -1),
make_tuple(16, 32, &vpx_sad16x32_msa, -1),
make_tuple(16, 16, &vpx_sad16x16_msa, -1),
make_tuple(16, 8, &vpx_sad16x8_msa, -1),
make_tuple(8, 16, &vpx_sad8x16_msa, -1),
make_tuple(8, 8, &vpx_sad8x8_msa, -1),
make_tuple(8, 4, &vpx_sad8x4_msa, -1),
make_tuple(4, 8, &vpx_sad4x8_msa, -1),
make_tuple(4, 4, &vpx_sad4x4_msa, -1),
};
INSTANTIATE_TEST_CASE_P(MSA, SADTest, ::testing::ValuesIn(msa_tests));
const SadMxNAvgParam avg_msa_tests[] = {
make_tuple(64, 64, &vpx_sad64x64_avg_msa, -1),
make_tuple(64, 32, &vpx_sad64x32_avg_msa, -1),
make_tuple(32, 64, &vpx_sad32x64_avg_msa, -1),
make_tuple(32, 32, &vpx_sad32x32_avg_msa, -1),
make_tuple(32, 16, &vpx_sad32x16_avg_msa, -1),
make_tuple(16, 32, &vpx_sad16x32_avg_msa, -1),
make_tuple(16, 16, &vpx_sad16x16_avg_msa, -1),
make_tuple(16, 8, &vpx_sad16x8_avg_msa, -1),
make_tuple(8, 16, &vpx_sad8x16_avg_msa, -1),
make_tuple(8, 8, &vpx_sad8x8_avg_msa, -1),
make_tuple(8, 4, &vpx_sad8x4_avg_msa, -1),
make_tuple(4, 8, &vpx_sad4x8_avg_msa, -1),
make_tuple(4, 4, &vpx_sad4x4_avg_msa, -1),
};
INSTANTIATE_TEST_CASE_P(MSA, SADavgTest, ::testing::ValuesIn(avg_msa_tests));
const SadMxNx4Param x4d_msa_tests[] = {
make_tuple(64, 64, &vpx_sad64x64x4d_msa, -1),
make_tuple(64, 32, &vpx_sad64x32x4d_msa, -1),
make_tuple(32, 64, &vpx_sad32x64x4d_msa, -1),
make_tuple(32, 32, &vpx_sad32x32x4d_msa, -1),
make_tuple(32, 16, &vpx_sad32x16x4d_msa, -1),
make_tuple(16, 32, &vpx_sad16x32x4d_msa, -1),
make_tuple(16, 16, &vpx_sad16x16x4d_msa, -1),
make_tuple(16, 8, &vpx_sad16x8x4d_msa, -1),
make_tuple(8, 16, &vpx_sad8x16x4d_msa, -1),
make_tuple(8, 8, &vpx_sad8x8x4d_msa, -1),
make_tuple(8, 4, &vpx_sad8x4x4d_msa, -1),
make_tuple(4, 8, &vpx_sad4x8x4d_msa, -1),
make_tuple(4, 4, &vpx_sad4x4x4d_msa, -1),
};
INSTANTIATE_TEST_CASE_P(MSA, SADx4Test, ::testing::ValuesIn(x4d_msa_tests));
#endif // HAVE_MSA
} // namespace
Reference in New Issue View Git Blame Copy Permalink