blob: 08dcef7a8ec4c096faa934bd3ca261c0ed54ccf9 [file] [log] [blame]
// Copyright 2021 Google LLC
//
// This source code is licensed under the BSD-style license found in the
// LICENSE file in the root directory of this source tree.
#include <xnnpack/AlignedAllocator.h>
#include <xnnpack/common.h>
#include <xnnpack/params.h>
#include <xnnpack/transpose.h>
#include <algorithm>
#include <cmath>
#include <functional>
#include <numeric>
#include <vector>
#include "bench/utils.h"
#include <benchmark/benchmark.h>
void transpose(
benchmark::State& state,
xnn_x8_transpose_ukernel_function transpose,
benchmark::utils::IsaCheckFunction isa_check = nullptr)
{
if (isa_check && !isa_check(state)) {
return;
}
const size_t height = state.range(0);
const size_t width = state.range(1);
const size_t tile_hbytes = height * sizeof(uint8_t);
const size_t tile_wbytes = width * sizeof(uint8_t);
std::vector<uint8_t, AlignedAllocator<uint8_t, 64>> x(
height * width + XNN_EXTRA_BYTES / sizeof(uint8_t));
std::vector<uint8_t, AlignedAllocator<uint8_t, 64>> y(
height * width + XNN_EXTRA_BYTES / sizeof(uint8_t));
std::iota(x.begin(), x.end(), 0);
std::fill(y.begin(), y.end(), 0);
for (auto _ : state) {
transpose(x.data(), y.data(), tile_wbytes, tile_hbytes, width,
height);
}
const uint64_t cpu_frequency = benchmark::utils::GetCurrentCpuFrequency();
if (cpu_frequency != 0) {
state.counters["cpufreq"] = cpu_frequency;
}
}
static void BenchmarkKernelSize(benchmark::internal::Benchmark* b)
{
b->ArgNames({"height", "width"});
b->Args({32, 32});
b->Args({64, 64});
b->Args({117, 117});
b->Args({1024, 1024});
}
BENCHMARK_CAPTURE(transpose, 1x2_scalar_int, xnn_x8_transpose_ukernel__1x2_scalar_int)
->Apply(BenchmarkKernelSize)->UseRealTime();
BENCHMARK_CAPTURE(transpose, 1x4_scalar_int, xnn_x8_transpose_ukernel__1x4_scalar_int)
->Apply(BenchmarkKernelSize)->UseRealTime();
BENCHMARK_CAPTURE(transpose, 2x1_scalar_int, xnn_x8_transpose_ukernel__2x1_scalar_int)
->Apply(BenchmarkKernelSize)->UseRealTime();
BENCHMARK_CAPTURE(transpose, 2x2_scalar_int, xnn_x8_transpose_ukernel__2x2_scalar_int)
->Apply(BenchmarkKernelSize)->UseRealTime();
BENCHMARK_CAPTURE(transpose, 2x4_scalar_int, xnn_x8_transpose_ukernel__2x4_scalar_int)
->Apply(BenchmarkKernelSize)->UseRealTime();
BENCHMARK_CAPTURE(transpose, 4x1_scalar_int, xnn_x8_transpose_ukernel__4x1_scalar_int)
->Apply(BenchmarkKernelSize)->UseRealTime();
BENCHMARK_CAPTURE(transpose, 4x2_scalar_int, xnn_x8_transpose_ukernel__4x2_scalar_int)
->Apply(BenchmarkKernelSize)->UseRealTime();
BENCHMARK_CAPTURE(transpose, 4x4_scalar_int, xnn_x8_transpose_ukernel__4x4_scalar_int)
->Apply(BenchmarkKernelSize)->UseRealTime();
#if XNN_ARCH_ARM || XNN_ARCH_ARM64
BENCHMARK_CAPTURE(transpose, 16x16_reuse_mov_zip_neon, xnn_x8_transpose_ukernel__16x16_reuse_mov_zip_neon)
->Apply(BenchmarkKernelSize)->UseRealTime();
BENCHMARK_CAPTURE(transpose, 16x16_reuse_switch_neon, xnn_x8_transpose_ukernel__16x16_reuse_switch_zip_neon)
->Apply(BenchmarkKernelSize)->UseRealTime();
#endif // XNN_ARCH_ARM || XNN_ARCH_ARM64
#if XNN_ARCH_X86 || XNN_ARCH_X86_64
BENCHMARK_CAPTURE(transpose, 16x16_reuse_mov_sse2, xnn_x8_transpose_ukernel__16x16_reuse_mov_sse2)
->Apply(BenchmarkKernelSize)->UseRealTime();
BENCHMARK_CAPTURE(transpose, 16x16_reuse_switch_sse2, xnn_x8_transpose_ukernel__16x16_reuse_switch_sse2)
->Apply(BenchmarkKernelSize)->UseRealTime();
#endif // XNN_ARCH_X86 || XNN_ARCH_X86_64
#ifndef XNNPACK_BENCHMARK_NO_MAIN
BENCHMARK_MAIN();
#endif