Skip to content

dealii: add constraint on suite-sparse for old gcc#9026

Merged
adamjstewart merged 1 commit intospack:developfrom
davydden:pkg/dealii_suite_sparse
Aug 20, 2018
Merged

dealii: add constraint on suite-sparse for old gcc#9026
adamjstewart merged 1 commit intospack:developfrom
davydden:pkg/dealii_suite_sparse

Conversation

@davydden
Copy link
Copy Markdown
Member

No description provided.

Copy link
Copy Markdown
Member

@adamjstewart adamjstewart left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pretty sure you need a space after the comma.

@davydden davydden force-pushed the pkg/dealii_suite_sparse branch from 1df7575 to f54b13e Compare August 20, 2018 14:42
@adamjstewart adamjstewart merged commit 271633e into spack:develop Aug 20, 2018
anderson2981 pushed a commit to anderson2981/spack that referenced this pull request Sep 7, 2018
@balay
Copy link
Copy Markdown
Contributor

balay commented Oct 3, 2018

I'm having an issue with this

[balay@cooleylogin2 spack]$ spack spec dealii
Input spec
--------------------------------
dealii

Concretized
--------------------------------
==> Error: An unsatisfiable version constraint has been detected for spec:

    [email protected]%[email protected]~cuda~openmp+pic~tbb arch=linux-rhel7-x86_64 


while trying to concretize the partial spec:

    [email protected]%[email protected]+adol-c+arpack+assimp build_type=DebugRelease ~cuda cuda_arch= ~doc+gmsh+gsl+hdf5~int64+metis+mpi+nanoflann+netcdf+oce~optflags+p4est+petsc~python+scalapack+slepc+sundials+trilinos arch=linux-rhel7-x86_64 
        ^[email protected]:
        ^arpack-ng+mpi
            ^mpi
            ^[email protected]%[email protected] build_type=RelWithDebInfo ~external-blas+lapacke+shared~xblas arch=linux-rhel7-x86_64 
        ^assimp
            ^[email protected]:1.63,1.65.1,1.67.0:+iostreams+serialization+system+thread
                ^bzip2
                ^[email protected]%[email protected]+optimize+pic+shared arch=linux-rhel7-x86_64 
            ^[email protected]%[email protected]~doc+ncurses+openssl+ownlibs~qt arch=linux-rhel7-x86_64 
        ^gmsh+netgen+oce+tetgen
            ^gmp
                ^autoconf
                    ^[email protected]:
                    ^perl
                        ^gdbm
                            ^readline
                                ^ncurses
                                    ^pkgconfig
                ^automake
                ^[email protected]:
            ^netgen
            ^oce
            ^tetgen
        ^[email protected]:+hl+mpi
        ^intel-tbb@2019%[email protected] cxxstd=default +shared+tm arch=linux-rhel7-x86_64 
        ^metis@5:~int64+real64
        ^[email protected]%[email protected] arch=linux-rhel7-x86_64 
        ^nanoflann
        ^netcdf+mpi
        ^p4est
        ^sundials~pthread
        ^trilinos+amesos+aztec+epetra+ifpack+ml+muelu+sacado+teuchos
            ^glm
            ^matio


dealii requires suite-sparse version :5.1.0, but spec asked for 5.3.0

This goes away with:

diff --git a/var/spack/repos/builtin/packages/dealii/package.py b/var/spack/repos/builtin/packages/dealii/package.py
index ec6b537..472c7a8 100644
--- a/var/spack/repos/builtin/packages/dealii/package.py
+++ b/var/spack/repos/builtin/packages/dealii/package.py
@@ -122,7 +122,7 @@ class Dealii(CMakePackage, CudaPackage):
     depends_on('bzip2', when='@:8.99')
     depends_on('lapack')
     depends_on('muparser')
-    depends_on('suite-sparse')
+    #depends_on('suite-sparse')
     depends_on('suite-sparse@:5.1.0', when='%gcc@:4.8.99')
     depends_on('tbb')
     depends_on('zlib')

I guess this is a usecase for if gcc<4.9: use suite-sparse@:5.1.0 else: use suite-sparse but I don't know the spack syntax for it

@davydden
Copy link
Copy Markdown
Member Author

davydden commented Oct 3, 2018

I guess this is a usecase for if gcc<4.9: use suite-sparse@:5.1.0 else: use suite-sparse but I don't know the spack syntax for it

Yes, that is how this should ideally work but I don’t think there is any other syntax for this. Meaning, if it’s broken we need to come up with something else. I guess it’s a variation of this issue
#1781
that ocncretizer is too quick to settle on the most recent version without considering all other constrains.
@tgamblin is there any quick fix you can think of?

In the absence of other options, we could make 5.1.0 preferred version and remove the constraint from deal. As mentioned in
#9025 this does not have anything to do with deal.ii, but rather current limitations of concretizer

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants