[petsc-dev] mpi/cuda issue

[Barry Smith]

> On Mar 22, 2016, at 1:35 PM, Satish Balay <balay at mcs.anl.gov> wrote:
> 
> An alternative is to have another nightly build with
> --download-superlu_dist-commit=origin/master. [similary add in other
> packages that currenlty use origin/master]

  This could be ok. Perhaps have one packages test that uses master for all the external packages possible?

   Barry

> 
> This test is to detect issues early with any upstream changes.
> 
> The update in master would be manual process anyway.. [so one can do
> it manually via next (after full tests)..]
> 
> Satish
> 
> On Tue, 22 Mar 2016, Barry Smith wrote:
> 
>> 
>>   We can have next track the master of superlu_dist and then update the commit on master when the next is ok after a change to superlu_dist. If we track purely manually we will be months behind.
>> 
>>   Barry
>> 
>>> On Mar 22, 2016, at 12:11 PM, Satish Balay <balay at mcs.anl.gov> wrote:
>>> 
>>> On Mon, 21 Mar 2016, Barry Smith wrote:
>>> 
>>>> 
>>>>> On Mar 21, 2016, at 10:04 PM, Satish Balay <balay at mcs.anl.gov> wrote:
>>>>> 
>>>>> On Mon, 21 Mar 2016, Barry Smith wrote:
>>>>> 
>>>>>> 
>>>>>>> On Mar 21, 2016, at 9:50 PM, Satish Balay <balay at mcs.anl.gov> wrote:
>>>>>>> 
>>>>>>> BTW: perils of using 'gitcommit=origin/master'
>>>>>>> http://ftp.mcs.anl.gov/pub/petsc/nightlylogs/archive/2016/03/21/master.html
>>>>>>> 
>>>>>>> Perhaps we should switch superlu_dist to use a working snapshot?
>>>>>>> self.gitcommit        = '35c3b21630d93b3f8392a68e607467c247b5e053'
>>>>>>> 
>>>>>>> balay at asterix /home/balay/petsc (master=)
>>>>>>> $ git grep origin/master config
>>>>>>> config/BuildSystem/config/packages/Chombo.py:    self.gitcommit        = 'origin/master'
>>>>>>> config/BuildSystem/config/packages/SuperLU_DIST.py:    self.gitcommit        = 'origin/master'
>>>>>>> config/BuildSystem/config/packages/amanzi.py:    self.gitcommit        = 'origin/master'
>>>>>>> config/BuildSystem/config/packages/saws.py:    self.gitcommit = 'origin/master'
>>>>>> 
>>>>>> Satish and Hong,
>>>>>> 
>>>>>> Sherry has changed SuperLU_dist to have no name conflicts with SuperLU this means we need to update the SuperLU_dist interface for fix these problems. Once things have settled down we can use a release commit instead of master.
>>>>> 
>>>>> Well the current issue is a breakage in superlu_dist build... [not the petsc interface]
>>>> 
>>>> Hmm, ok then we'll need to take a look at it.
>>> 
>>> I think its best to keep petsc master stable [wrt superlu_dist] - so
>>> changed master to use 35c3b21630d93b3f8392a68e607467c247b5e053
>>> 
>>> We should sync with superlu_dist manually [test in next - and then
>>> update master] - for any new changes..
>>> 
>>> Satish
>> 
>> 
>