mirror/eigen
2
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-03-07 18:27:40 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix flaky test for tensor fft.

Browse Source
This commit is contained in:
Rasmus Munk Larsen 2019-01-16 14:03:12 -08:00
parent 0f028f61cb
commit ee550a2ac3
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
unsupported/test/cxx11_tensor_fft.cpp
View File

@ -228,9 +228,6 @@ template <typename RealScalar>
static void test_fft_non_power_of_2_round_trip(int exponent) { static void test_fft_non_power_of_2_round_trip(int exponent) {
int n = (1 << exponent) + 1; int n = (1 << exponent) + 1;
// The dimension type needs to be at least 8 bytes long for the
// Tensor constructor to work. On Windows, long is only 4 bytes long,
// so use long long here to force the usage of a 8 bytes integer type.
Eigen::DSizes<std::int64_t, 1> dimensions; Eigen::DSizes<std::int64_t, 1> dimensions;
dimensions[0] = n; dimensions[0] = n;
const DSizes<std::int64_t, 1> arr = dimensions; const DSizes<std::int64_t, 1> arr = dimensions;
@ -249,7 +246,9 @@ static void test_fft_non_power_of_2_round_trip(int exponent) {
forward.template fft<RealPart, FFT_REVERSE>(fft); forward.template fft<RealPart, FFT_REVERSE>(fft);
for (int i = 0; i < n; ++i) { for (int i = 0; i < n; ++i) {
VERIFY_IS_APPROX(input[i], output[i]); RealScalar tol = test_precision<RealScalar>() *
(std::abs(input[i]) + std::abs(output[i]) + 1);
VERIFY_IS_APPROX_OR_LESS_THAN(std::abs(input[i] - output[i]), tol);
} }
} }
@ -301,4 +300,5 @@ EIGEN_DECLARE_TEST(cxx11_tensor_fft) {
test_fft_real_input_energy<RowMajor, double, false, Eigen::BothParts, FFT_FORWARD, 4>(); test_fft_real_input_energy<RowMajor, double, false, Eigen::BothParts, FFT_FORWARD, 4>();
test_fft_non_power_of_2_round_trip<float>(7); test_fft_non_power_of_2_round_trip<float>(7);
test_fft_non_power_of_2_round_trip<double>(7);
} }