blob: fc2373c94ed8d89097e6f0163070921773ea395c [file] [log] [blame]
// Copyright 2019 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.
#pragma once
#include <gtest/gtest.h>
#include <algorithm>
#include <cassert>
#include <cstddef>
#include <cstdlib>
#include <functional>
#include <random>
#include <vector>
#include <xnnpack.h>
#include <xnnpack/params.h>
class RAddExpMinusMaxMicrokernelTester {
public:
inline RAddExpMinusMaxMicrokernelTester& n(size_t n) {
assert(n != 0);
this->n_ = n;
return *this;
}
inline size_t n() const {
return this->n_;
}
inline RAddExpMinusMaxMicrokernelTester& iterations(size_t iterations) {
this->iterations_ = iterations;
return *this;
}
inline size_t iterations() const {
return this->iterations_;
}
void test(xnn_f32_raddexpminusmax_ukernel_function raddexpminusmax) const {
std::random_device random_device;
auto rng = std::mt19937(random_device());
// Choose such range that expf(x[i]) overflows, but expf(x[i] - x_max) doesn't.
// However, the range is still narrow enough that double-precision exp doesn't overflow.
auto f32rng = std::bind(std::uniform_real_distribution<float>(90.0f, 100.0f), rng);
std::vector<float> x(n() + XNN_EXTRA_BYTES / sizeof(float));
for (size_t iteration = 0; iteration < iterations(); iteration++) {
std::generate(x.begin(), x.end(), std::ref(f32rng));
// Compute reference results.
double sum_ref = 0.0f;
const float x_max = *std::max_element(x.begin(), x.begin() + n());
for (size_t i = 0; i < n(); i++) {
sum_ref += exp(x[i] - x_max);
}
// Call optimized micro-kernel.
float sum = std::nanf("");
raddexpminusmax(n() * sizeof(float), x.data(), &sum, x_max);
// Verify results.
ASSERT_NEAR(sum_ref, double(sum), std::abs(sum_ref) * 1.0e-6)
<< "n = " << n() << ", x_max = " << x_max;
}
}
private:
size_t n_{1};
size_t iterations_{15};
};