You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It has been reported that the tests fail when using the nFortran (from nvidia) compiler. One is a rounding error difference. Another is that parsing the control file for the SplineGeometry mesh fails with exceptions saying that blocks are not properly terminated. (\end{...} does not contain the proper keyword.) This is not true by visual inspection of the control file, nor does gFortran have a problem with it. As far as I can tell nFortran only runs on linux. Here are the error messages:
////////////////////////////////
//// Begin Test Suites ////
////////////////////////////////
-------------------------------------------------------------
Summary of failed tests for test suite: 0 alloc: need 128 size 160 p 0x10fbb70 area 0x10fbb80 end 0x10fbc0f
Curve evaluation tests0 dealloc p 0x488720 area 0x10fbb80
0 failures out of 13 assertions.
-------------------------------------------------------------
-------------------------------------------------------------
Summary of failed tests for test suite: 0 alloc: need 128 size 160 p 0x10fbb70 area 0x10fbb80 end 0x10fbc0f
Gaussian Curvature evaluation tests0 dealloc p 0x488720 area 0x10fbb80
1 failures out of 4 assertions.
-------------------------------------------------------------
Double Precision equality failed:
Quadratic Hessian error
Expected [0.000000000000000], Got [1.0049518550658831E-008]
-------------------------------------------------------------
Summary of failed tests for test suite: 0 alloc: need 128 size 160 p 0x10fbb70 area 0x10fbb80 end 0x10fbc0f
BiCubic Interpolation tests0 dealloc p 0x488720 area 0x10fbb80
0 failures out of 6 assertions.
-------------------------------------------------------------
-------------------------------------------------------------
Summary of failed tests for test suite: 0 alloc: need 128 size 160 p 0x177f780 area 0x177f790 end 0x177f81f
ControlFile: Benchmarks/ControlFiles/Pill.control0 dealloc p 0x488720 area 0x177f790
0 failures out of 9 assertions.
-------------------------------------------------------------
I think that the rounding error failure will be solved once #76 merges into main. It is strange that the file parsing is the issue as we now know that in addition to gfortran, ifort also has no issue with it.
It has been reported that the tests fail when using the nFortran (from nvidia) compiler. One is a rounding error difference. Another is that parsing the control file for the SplineGeometry mesh fails with exceptions saying that blocks are not properly terminated. (\end{...} does not contain the proper keyword.) This is not true by visual inspection of the control file, nor does gFortran have a problem with it. As far as I can tell nFortran only runs on linux. Here are the error messages:
////////////////////////////////
//// Begin Test Suites ////
////////////////////////////////
-------------------------------------------------------------
Summary of failed tests for test suite: 0 alloc: need 128 size 160 p 0x10fbb70 area 0x10fbb80 end 0x10fbc0f
Curve evaluation tests0 dealloc p 0x488720 area 0x10fbb80
Gaussian Curvature evaluation tests0 dealloc p 0x488720 area 0x10fbb80
BiCubic Interpolation tests0 dealloc p 0x488720 area 0x10fbb80
ControlFile: Benchmarks/ControlFiles/Pill.control0 dealloc p 0x488720 area 0x177f790
Improper block termination. Expected SPLINE_CURVE; found SPLINE_DATA
Block termination error. Expected: OUTER_BOUNDARY; Found: SPLINE_CURVE
The text was updated successfully, but these errors were encountered: