blob: 7a3067dade5d0074efbfbbe38868ca173d1a0098 [file] [log] [blame]
/*
* Copyright (c) 2018, Alliance for Open Media. All rights reserved
*
* This source code is subject to the terms of the BSD 2 Clause License and
* the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License
* was not distributed with this source code in the LICENSE file, you can
* obtain it at www.aomedia.org/license/software. If the Alliance for Open
* Media Patent License 1.0 was not distributed with this source code in the
* PATENTS file, you can obtain it at www.aomedia.org/license/patent.
*/
#include <tuple>
#include "third_party/googletest/src/googletest/include/gtest/gtest.h"
#include "aom/aom_integer.h"
#include "aom_ports/aom_timer.h"
#include "av1/encoder/ml.h"
#include "config/aom_config.h"
#include "config/aom_dsp_rtcd.h"
#include "config/av1_rtcd.h"
#include "test/util.h"
#include "test/register_state_check.h"
#include "test/acm_random.h"
namespace {
typedef void (*NnPredict_Func)(const float *const input_nodes,
const NN_CONFIG *const nn_config,
int reduce_prec, float *const output);
typedef std::tuple<const NnPredict_Func> NnPredictTestParam;
const float epsilon = 1e-3f; // Error threshold for functional equivalence
class NnPredictTest : public ::testing::TestWithParam<NnPredictTestParam> {
public:
virtual void SetUp() {
const int MAX_NODES2 = NN_MAX_NODES_PER_LAYER * NN_MAX_NODES_PER_LAYER;
// Allocate two massive buffers on the heap for edge weights and node bias
// Then set-up the double-dimension arrays pointing into the big buffers
weights_buf = (float *)aom_malloc(MAX_NODES2 * (NN_MAX_HIDDEN_LAYERS + 1) *
sizeof(*weights_buf));
bias_buf =
(float *)aom_malloc(NN_MAX_NODES_PER_LAYER *
(NN_MAX_HIDDEN_LAYERS + 1) * sizeof(*bias_buf));
ASSERT_NE(weights_buf, nullptr);
ASSERT_NE(bias_buf, nullptr);
for (int i = 0; i < NN_MAX_HIDDEN_LAYERS + 1; i++) {
weights[i] = &weights_buf[i * MAX_NODES2];
bias[i] = &bias_buf[i * NN_MAX_NODES_PER_LAYER];
}
target_func_ = GET_PARAM(0);
}
virtual void TearDown() {
aom_free(weights_buf);
aom_free(bias_buf);
}
void RunNnPredictTest(const NN_CONFIG *const shape);
void RunNnPredictSpeedTest(const NN_CONFIG *const shape, const int run_times);
void RunNnPredictTest_all(const NN_CONFIG *const shapes,
const int num_shapes);
void RunNnPredictSpeedTest_all(const NN_CONFIG *const shapes,
const int num_shapes, const int run_times);
private:
NnPredict_Func target_func_;
libaom_test::ACMRandom rng_;
float *weights[NN_MAX_HIDDEN_LAYERS + 1] = { 0 };
float *bias[NN_MAX_HIDDEN_LAYERS + 1] = { 0 };
float *weights_buf = nullptr, *bias_buf = nullptr;
};
GTEST_ALLOW_UNINSTANTIATED_PARAMETERIZED_TEST(NnPredictTest);
void NnPredictTest::RunNnPredictTest(const NN_CONFIG *const shape) {
float inputs[NN_MAX_NODES_PER_LAYER] = { 0 };
float outputs_test[NN_MAX_NODES_PER_LAYER] = { 0 };
float outputs_ref[NN_MAX_NODES_PER_LAYER] = { 0 };
NN_CONFIG nn_config;
memcpy(&nn_config, shape, sizeof(nn_config));
char shape_str[32] = { 0 };
snprintf(shape_str, sizeof(shape_str), "%d", shape->num_inputs);
for (int layer = 0; layer < shape->num_hidden_layers; layer++)
snprintf(&shape_str[strlen(shape_str)],
sizeof(shape_str) - strlen(shape_str), "x%d",
shape->num_hidden_nodes[layer]);
snprintf(&shape_str[strlen(shape_str)], sizeof(shape_str) - strlen(shape_str),
"x%d", shape->num_outputs);
for (int i = 0; i < NN_MAX_HIDDEN_LAYERS + 1; i++) {
nn_config.weights[i] = weights[i];
nn_config.bias[i] = bias[i];
}
for (int iter = 0; iter < 10000 && !HasFatalFailure(); ++iter) {
for (int node = 0; node < shape->num_inputs; node++) {
inputs[node] = ((float)rng_.Rand31() - (1 << 30)) / (1u << 31);
}
for (int layer = 0; layer < shape->num_hidden_layers; layer++) {
for (int node = 0; node < NN_MAX_NODES_PER_LAYER; node++) {
bias[layer][node] = ((float)rng_.Rand31() - (1 << 30)) / (1u << 31);
}
for (int node = 0; node < NN_MAX_NODES_PER_LAYER * NN_MAX_NODES_PER_LAYER;
node++) {
weights[layer][node] = ((float)rng_.Rand31() - (1 << 30)) / (1u << 31);
}
}
// Now the outputs:
int layer = shape->num_hidden_layers;
for (int node = 0; node < NN_MAX_NODES_PER_LAYER; node++) {
bias[layer][node] = ((float)rng_.Rand31() - (1 << 30)) / (1u << 31);
}
for (int node = 0; node < NN_MAX_NODES_PER_LAYER * NN_MAX_NODES_PER_LAYER;
node++) {
weights[layer][node] = ((float)rng_.Rand31() - (1 << 30)) / (1u << 31);
}
av1_nn_predict_c(inputs, &nn_config, 0, outputs_ref);
target_func_(inputs, &nn_config, 0, outputs_test);
for (int node = 0; node < shape->num_outputs; node++) {
if (outputs_ref[node] < epsilon) {
ASSERT_LE(outputs_test[node], epsilon)
<< "Reference output was near-zero, test output was not ("
<< shape_str << ")";
} else {
const float error = outputs_ref[node] - outputs_test[node];
const float relative_error = fabsf(error / outputs_ref[node]);
ASSERT_LE(relative_error, epsilon)
<< "Excessive relative error between reference and test ("
<< shape_str << ")";
}
}
}
}
void NnPredictTest::RunNnPredictSpeedTest(const NN_CONFIG *const shape,
const int run_times) {
float inputs[NN_MAX_NODES_PER_LAYER] = { 0 };
float outputs_test[NN_MAX_NODES_PER_LAYER] = { 0 };
float outputs_ref[NN_MAX_NODES_PER_LAYER] = { 0 };
NN_CONFIG nn_config;
memcpy(&nn_config, shape, sizeof(nn_config));
for (int i = 0; i < NN_MAX_HIDDEN_LAYERS; i++) {
nn_config.weights[i] = weights[i];
nn_config.bias[i] = bias[i];
}
// Don't bother actually changing the values for inputs/weights/bias: it
// shouldn't make any difference for a speed test.
aom_usec_timer timer;
aom_usec_timer_start(&timer);
for (int i = 0; i < run_times; ++i) {
av1_nn_predict_c(inputs, &nn_config, 0, outputs_ref);
}
aom_usec_timer_mark(&timer);
const double time1 = static_cast<double>(aom_usec_timer_elapsed(&timer));
aom_usec_timer_start(&timer);
for (int i = 0; i < run_times; ++i) {
target_func_(inputs, &nn_config, 0, outputs_test);
}
aom_usec_timer_mark(&timer);
const double time2 = static_cast<double>(aom_usec_timer_elapsed(&timer));
printf("%d", shape->num_inputs);
for (int layer = 0; layer < shape->num_hidden_layers; layer++)
printf("x%d", shape->num_hidden_nodes[layer]);
printf("x%d: ", shape->num_outputs);
printf("%7.2f/%7.2fns (%3.2f)\n", time1, time2, time1 / time2);
}
// This is all the neural network shapes observed executed in a few different
// runs of the encoder. It also conveniently covers all the kernels
// implemented.
static const NN_CONFIG shapes[] = {
{ 10, 16, 1, { 64 }, { 0 }, { 0 } }, { 12, 1, 1, { 12 }, { 0 }, { 0 } },
{ 12, 1, 1, { 24 }, { 0 }, { 0 } }, { 12, 1, 1, { 32 }, { 0 }, { 0 } },
{ 18, 4, 1, { 24 }, { 0 }, { 0 } }, { 18, 4, 1, { 32 }, { 0 }, { 0 } },
{ 4, 1, 1, { 16 }, { 0 }, { 0 } }, { 8, 1, 1, { 16 }, { 0 }, { 0 } },
{ 8, 4, 1, { 16 }, { 0 }, { 0 } }, { 8, 1, 1, { 24 }, { 0 }, { 0 } },
{ 8, 1, 1, { 32 }, { 0 }, { 0 } }, { 8, 1, 1, { 64 }, { 0 }, { 0 } },
{ 9, 3, 1, { 32 }, { 0 }, { 0 } }, { 4, 4, 1, { 8 }, { 0 }, { 0 } },
};
void NnPredictTest::RunNnPredictTest_all(const NN_CONFIG *const shapes,
const int num_shapes) {
for (int i = 0; i < num_shapes; i++) RunNnPredictTest(&shapes[i]);
}
void NnPredictTest::RunNnPredictSpeedTest_all(const NN_CONFIG *const shapes,
const int num_shapes,
const int run_times) {
for (int i = 0; i < num_shapes; i++)
NnPredictTest::RunNnPredictSpeedTest(&shapes[i], run_times);
}
TEST_P(NnPredictTest, RandomValues) {
RunNnPredictTest_all(shapes, sizeof(shapes) / sizeof(*shapes));
}
TEST_P(NnPredictTest, DISABLED_Speed) {
RunNnPredictSpeedTest_all(shapes, sizeof(shapes) / sizeof(*shapes), 10000000);
}
#if HAVE_SSE3 && !CONFIG_EXCLUDE_SIMD_MISMATCH
INSTANTIATE_TEST_SUITE_P(SSE3, NnPredictTest,
::testing::Values(av1_nn_predict_sse3));
#endif
#if HAVE_NEON
INSTANTIATE_TEST_SUITE_P(NEON, NnPredictTest,
::testing::Values(av1_nn_predict_neon));
#endif
} // namespace