[petsc-users] How can PETSc configure with MSMPI

[Jed Brown]

Satish Balay <balay at mcs.anl.gov> writes:
> We are not bundling it. We are using it for our nightly tests.

But are we recommending that people use it?

> Also the comparision to using petsc doesn't apply here. PETSc changes
> API with every release. As long as we say we support MPI-2 - such
> things shold work. In that sense - its good to have some testsuites
> with such old stuff.

Yes, testing in private is different from recommending that people use
it.

> Supporting windows is a mess. There is no single 'easily' supported
> thingy here.
>
> Sure we should probably also have a different test infracture with
> latest compilers, latest supported mpi etc.. But that doesn't mean
> testing with alternatives does not have value..
>
> 'gcc is now at 4.8 - so we shoud not support older versions of gcc...'

It would be different if we got a bug report with the latest compiler
and we told people to use an old compiler.  (Some DOE people have gotten
mileage out of that tactic, but we find it abhorrent.)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 835 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20140202/3f233d64/attachment.pgp>