Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Release 6.0.1? #106

Closed
gruenich opened this issue Jul 7, 2023 · 11 comments
Closed

Release 6.0.1? #106

gruenich opened this issue Jul 7, 2023 · 11 comments

Comments

@gruenich
Copy link
Contributor

gruenich commented Jul 7, 2023

@xiaoyeli, what do think about releasing the current master as 6.0.1?

The fixed regression #100 would justify the change. Having a proper release would help downstream projects like PETSc.

Further, some of the fixed compiler warnings might be turned into errors with upcoming compiler versions. The improvements to CMake and Doxygen come on top.

@wuch-g2v
Copy link

+1

@gruenich
Copy link
Contributor Author

gruenich commented Aug 6, 2023

Thank you for creating a 6.0.1 release! I am on my way to update the openSuse package.

Before I close this issue, could you please update the homepage and mention 6.0.1 there?

@kloczek
Copy link

kloczek commented Aug 6, 2023

Looks like it some issue with 6.0.1

[ 80%] Building C object EXAMPLE/CMakeFiles/citersol.dir/cfgmr.c.o
cd /home/tkloczko/rpmbuild/BUILD/superlu-6.0.1/x86_64-redhat-linux-gnu/EXAMPLE && /usr/bin/gcc  -I/home/tkloczko/rpmbuild/BUILD/superlu-6.0.1/x86_64-redhat-linux-gnu/SRC -I/home/tkloczko/rpmbuild/BUILD/superlu-6.0.1/SRC -DUSE_VENDOR_BLAS -O2 -g -grecord-gcc-switches -pipe -Wall -Werror=format-security -Wp,-U_FORTIFY_SOURCE,-D_FORTIFY_SOURCE=3 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -fno-omit-frame-pointer -mno-omit-leaf-frame-pointer -m64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -fdata-sections -ffunction-sections -flto=auto -flto-partition=none -DNDEBUG -std=c99 -MD -MT EXAMPLE/CMakeFiles/citersol.dir/cfgmr.c.o -MF CMakeFiles/citersol.dir/cfgmr.c.o.d -o CMakeFiles/citersol.dir/cfgmr.c.o -c /home/tkloczko/rpmbuild/BUILD/superlu-6.0.1/EXAMPLE/cfgmr.c
/home/tkloczko/rpmbuild/BUILD/superlu-6.0.1/EXAMPLE/cfgmr.c: In function ‘cfgmr’:
/home/tkloczko/rpmbuild/BUILD/superlu-6.0.1/EXAMPLE/cfgmr.c:83:16: error: conflicting types for ‘caxpy_’; have ‘int(int *, complex *, complex *, int *, complex *, int *)’
   83 |     extern int caxpy_(int *, complex *, complex [], int *, complex [], int *);
      |                ^~~~~~
/home/tkloczko/rpmbuild/BUILD/superlu-6.0.1/EXAMPLE/cfgmr.c:38:13: note: previous declaration of ‘caxpy_’ with type ‘void(int *, complex *, complex *, int *, complex *, int *)’
   38 | extern void caxpy_(int *, complex *, complex *, int *, complex *, int *);
      |             ^~~~~~
make[2]: *** [EXAMPLE/CMakeFiles/citersol.dir/build.make:93: EXAMPLE/CMakeFiles/citersol.dir/cfgmr.c.o] Error 1

@wo80
Copy link
Contributor

wo80 commented Aug 6, 2023

@kloczek Which BLAS provider are you using? If you intended to use the internal CBLAS provided by SuperLU, have you set enable_internal_blaslib?

@kloczek
Copy link

kloczek commented Aug 6, 2023

flexiblas 3.3.1

@wo80
Copy link
Contributor

wo80 commented Aug 6, 2023

@kloczek Did previous versions of SuperLU compile using flexiblas? I see some changes affecting the prototypes in cfgmr in 332cf3e , but that was before the release of 6.0.0. ...

@kloczek
Copy link

kloczek commented Aug 6, 2023

@kloczek Did previous versions of SuperLU compile using flexiblas? I see some changes affecting the cfgmr prototype in 332cf3e , but that was before the release of 6.0.0. ...

Just checked my build infra build logs (generated At 22 Jun) which stores record of correctly build 6.0.0 packages.

@wo80
Copy link
Contributor

wo80 commented Aug 6, 2023

I think the problem might be the specified return type void instead of int of caxpy_ in citersol. Though I couldn't find a commit that affected the prototype between 6.0.0 and 6.0.1. Other examples using *axpy_ might also be affected.

EDIT: see comment below ...

@wo80
Copy link
Contributor

wo80 commented Aug 6, 2023

Double checked void should be correct. So the return type int in cfgmr is wrong, should be void.

superlu/EXAMPLE/cfgmr.c

Lines 79 to 82 in d6a031c

/* prototypes */
extern int ccopy_(int *, complex *, int *, complex *, int *);
extern int caxpy_(int *, complex *, complex [], int *, complex [], int *);

EDIT:
The duplicate but correct definition was removed in commit 5ad0fe1 - after release of 6.0.0

@xiaoyeli
Copy link
Owner

xiaoyeli commented Aug 11, 2023 via email

@gruenich
Copy link
Contributor Author

Thank you for updating the documentation and the change log. The page itself still mentions v6.0.0 to be the latest... I am closing this issue, as I am fine with it.

The issue @kloczek brought up is discussed in #113, so no reason to keep this issue open.

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

No branches or pull requests

5 participants