devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix TensorReduction warnings and error bound for sum accuracy test.

Browse Source 
The sum accuracy test currently uses the default test precision for
the given scalar type.  However, scalars are generated via a normal
distribution, and given a large enough count and strong enough random
generator, the expected sum is zero.  This causes the test to
periodically fail.

Here we estimate an upper-bound for the error as `sqrt(N) * prec` for
summing N values, with each having an approximate epsilon of `prec`.

Also fixed a few warnings generated by MSVC when compiling the
reduction test.
This commit is contained in:
Antonio Sanchez
2021-10-29 22:03:44 +00:00
committed by Antonio Sanchez
parent b3bea43a2d
commit f6c8cc0e99
3 changed files with 8 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
unsupported/test/cxx11_tensor_reduction.cpp
View File

@@ -504,7 +504,11 @@ void test_sum_accuracy() {
for (int i = 0; i < num_elements; ++i) {
expected_sum += static_cast<double>(tensor(i));
}
VERIFY_IS_APPROX(sum(), static_cast<ScalarType>(expected_sum));
// Scale tolerance to account for # elements. Otherwise, we periodically fail, since
// E[sum] == prescribed_mean == 0 for the first iteration.
double err = Eigen::numext::abs(static_cast<double>(sum()) - expected_sum);
double tol = Eigen::numext::sqrt(num_elements) * static_cast<double>(test_precision<ScalarType>()) * numext::maxi(1.0, prescribed_mean);
VERIFY(err < tol);
}
}