GitHub-Proxy/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2025-06-04 18:54:00 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix wrong negative in nullary unit test when extended precision is used (FPU).

Browse Source
This commit is contained in:
Gael Guennebaud 2014-12-16 16:23:47 +01:00
parent 7dad5f797e
commit 99501a2c4c
1 changed files with 3 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
test/nullary.cpp
View File

@ -80,7 +80,9 @@ void testVectorType(const VectorType& base)
Matrix<Scalar,1,Dynamic> col_vector(size); Matrix<Scalar,1,Dynamic> col_vector(size);
row_vector.setLinSpaced(size,low,high); row_vector.setLinSpaced(size,low,high);
col_vector.setLinSpaced(size,low,high); col_vector.setLinSpaced(size,low,high);
VERIFY( row_vector.isApprox(col_vector.transpose(), NumTraits<Scalar>::epsilon())); // when using the extended precision (e.g., FPU) the relative error might exceed 1 bit
// when computing the squared sum in isApprox, thus the 2x factor.
VERIFY( row_vector.isApprox(col_vector.transpose(), Scalar(2)*NumTraits<Scalar>::epsilon()));
Matrix<Scalar,Dynamic,1> size_changer(size+50); Matrix<Scalar,Dynamic,1> size_changer(size+50);
size_changer.setLinSpaced(size,low,high); size_changer.setLinSpaced(size,low,high);