Closed mattkab2 closed 4 years ago
General comment is to make sure line continuations have 4 spaces.
END DO
-> ENDDO
END IF
-> ENDIF
Note: I plan to squash some of these commits (doxygen, whitespace, etc) together before the merge eventually happens. Makes no sense to have 5 commits dedicated to fixing comments/whitespace.
Note: I plan to squash some of these commits (doxygen, whitespace, etc) together before the merge eventually happens. Makes no sense to have 5 commits dedicated to fixing comments/whitespace.
just squash merge the whole thing when the time comes
@mattkab2 You'll need to rebase/merge in master due to the change in indentation PR.
@bscollin and @stimpsonsg This is all set. I'd like to click merge, but wanted to check with you both first. The only thing I was thinking was if you had concerns with the Trilinos merge. There shouldn't be any overlap/conflict with that PR.
@bkochuna The only thing left I wanted to do (which Dan reminded me of after the meeting) was to double check the changes don't affect MPACT tests. I checked the parallel builds, which all passed, and was going to check the serial ones tomorrow. I'll attach the logs when they complete.
As promised, here are the ctest logs from an MPACT build using this Futility branch: ctest_serBuild.log ctest_parBuild.log Note there is one failing test in the parallel build because one of the nightly tests requires COBRATF which was not included in the build.
So all clear to merge?
@bkochuna Yes, all clear
Thanks @mattkab2 but I meant @bscollin
Alright. I'm merging today at noon as I've not heard back, and there were no definitive reasons to hold this up any longer.
Looks like this broke mpi builds without PETSc:
#ifdef FUTILITY_HAVE_PETSC
#include <petscversion.h>
#if ((PETSC_VERSION_MAJOR>=3) && (PETSC_VERSION_MINOR>=6))
#include <petsc/finclude/petsc.h>
#else
#include <finclude/petsc.h>
#endif
#undef IS
#endif
Missing an INCLUDE 'mpif.h'
if PETSc isn't enabled.
@sooyoung0321 or @mattkab2 can one of you address?
@sooyoung0321 or @mattkab2 can one of you address? I'll take a look.
I've updated NativeParallelDatatypes branch but it does not appear in this pull request.
@sooyoung0321 Make the change on master, push it to a new branch, and create a new pull request. This pull request has been merged and closed, so you shouldn't push the changes into this branch/pull request.
This branch is the first in a series of 3, split off from the bandMat_pGMRES_rsor branch.
These changes pertain to the creation of new parallel datatypes (3 banded matrix types and 1 distributed vector type) with the associated MatVec/dot/norm routines.