[petsc-dev] petsc patch number

Kong, Fande fande.kong at inl.gov
Wed Apr 19 11:06:46 CDT 2017


On Wed, Apr 19, 2017 at 9:59 AM, Satish Balay <balay at mcs.anl.gov> wrote:

> Why do you need to distinguish (between petsc maint snapshots) if you
> are using the fixed superlu_dist version with both?
>

The old one has been complied several months ago, and we do not want to
recompile it, and most of tests stay with this version. The current patched
one is used for a few tests only.  It is not that trivial for me to update
all user side applications with petsc. We therefore avoid updating petsc
frequently.

 Fande,


> Satish
>
> On Wed, 19 Apr 2017, Kong, Fande wrote:
>
> > Thanks, Satish,
> >
> > Yes, but the petsc version is till 3.7.5, right? If so, according to the
> > petsc version number,  how can I tell differences between this petsc and
> > the old petsc whose version is also 3.7.5.?
> >
> >
> > Fande,
> >
> > On Wed, Apr 19, 2017 at 9:40 AM, Satish Balay <balay at mcs.anl.gov> wrote:
> >
> > > Use --download-superlu_dist=url-to-fixed-superlu_dist-version
> > >
> > > [or if using superlu_dist from git]
> > > --download-superlu_dist-commit=commit
> > >
> > > Current maint has points to 'v5.1.3' - so you could use:
> > > --download-superlu_dist-commit=v5.1.3
> > >
> > > Satish
> > >
> > > On Wed, 19 Apr 2017, Kong, Fande wrote:
> > >
> > > > Thanks for your email.  Sorry for late reply.
> > > >
> > > > On Thu, Apr 13, 2017 at 10:20 PM, Jed Brown <jed at jedbrown.org>
> wrote:
> > > >
> > > > > Individual commits are not strictly ordered.  'maint' does not get
> API
> > > > > changes.  What are you trying to test for?  Presence of a bug fix?
> > > > >
> > > >
> > > > Yes, I am going to test for the bug fix.
> > > >
> > > > There is a test system on the moose-side. We usually test all
> examples
> > > > against one or two versions of PETSc, (PETSc-3.7.5 and/or
> PETSc-3.6.4,
> > > > right now). The PETSc is precompiled and sits in the test system.
> Now,
> > > > there is a bug fix for superlu_dist in the current patched petsc. We
> want
> > > > to test SOME of examples (not all of them) against the current
> patched
> > > > petsc (the version number is still the same, 3.7.5). In a test file,
> we
> > > > have something like "petsc_version= x.x.x " to tell the test system
> which
> > > > version of petsc we want to use.  But now, I can not distinguish the
> > > petsc
> > > > with the broken superlu_dist and the current patched petsc by using
> the
> > > > petsc version number.
> > > >
> > > > Any suggestions?
> > > >
> > > > Fande,
> > > >
> > > >
> > > >
> > > > >
> > > > > "Kong, Fande" <fande.kong at inl.gov> writes:
> > > > >
> > > > > > Thanks,
> > > > > >
> > > > > > In the moose side, we want to make some tests against the current
> > > maint,
> > > > > > but still have some tests against the old one. The two versions
> of
> > > petsc
> > > > > > may have the same petsc version number, e.g. 3.7.5. I want a
> patch
> > > number
> > > > > > from which I could tell which one is newer. How it is different
> from
> > > the
> > > > > > old one.
> > > > > >
> > > > > > Any alternative way to do that? When compiling  PETSc, do we
> have a
> > > macro
> > > > > > variable to define the commit number?
> > > > > >
> > > > > > Fande,
> > > > > >
> > > > > > On Thu, Apr 13, 2017 at 12:06 PM, Satish Balay <
> balay at mcs.anl.gov>
> > > > > wrote:
> > > > > >
> > > > > >> On Thu, 13 Apr 2017, Kong, Fande wrote:
> > > > > >>
> > > > > >> > Hi Developers,
> > > > > >> >
> > > > > >> > Do we increase the patch number when pushing a new patch into
> > > maint?
> > > > > >>
> > > > > >> nope
> > > > > >>
> > > > > >> > If not, how to know if the current maint is different from the
> > > one I
> > > > > >> > just got yesterday. I know a commit number is always
> available.
> > > > > >>
> > > > > >> git fetch && git log maint..origin/maint
> > > > > >>
> > > > > >> [and you can look at that and decide to do a 'git pull' on
> maint]
> > > > > >>
> > > > > >> Satish
> > > > > >>
> > > > >
> > > >
> > >
> > >
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20170419/515be59a/attachment.html>


More information about the petsc-dev mailing list