Skip to content

Conversation

@martin-frbg
Copy link
Collaborator

second part of workaround for #2709

@mattip
Copy link
Contributor

mattip commented Oct 12, 2020

My local build failed, any ideas ?

@martin-frbg
Copy link
Collaborator Author

Looks like you are missing the

if (DEFINED HAVE_SSE3)
    set (KERNEL_DEFINITIONS "${KERNEL_DEFINITIONS} -msse3")
 endif()

addition to cmake/system.cmake (bb74dd2 - sorry for not going through the PR mechanism with that one but it was a late night fix for a merge conflict)

@mattip
Copy link
Contributor

mattip commented Oct 12, 2020

No cookie. I still get the same errors as in #2882. Is there a more systematic way to do this?

@martin-frbg
Copy link
Collaborator Author

hm strange. The other fpu-using .S files are all unused AFAICT, perhaps the uncleared exception is now triggered by something in gfortran's compilation of the lapack code.

@martin-frbg
Copy link
Collaborator Author

btw this PR is likely to be obsolete as soon as #2888 gets merged. Now all we need is a decent reimplementation of nrm2 and OpenBLAS is going to be just another hapless victim of the Windows bug. :-/

@martin-frbg martin-frbg reopened this Oct 12, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants