[petsc-dev] 3rd party GPU AMG solvers

[Mark Adams]

On Fri, Jul 14, 2017 at 8:02 AM, Matthew Knepley <knepley at gmail.com> wrote:

> On Fri, Jul 14, 2017 at 6:53 AM, Mark Adams <mfadams at lbl.gov> wrote:
>
>> Karli, this would be great if you could investigate this.
>>
>> A lot of this is driven by desires of DOE programs -- not your monkey not
>> your circus -- but I think that we need to have a story for how to use
>> GPUs, or whatever apps in our funding community want to do, and tell it
>> dispassionately. We don't commit resources nor commit to design changes but
>> just say this how one could do it.
>>
>
> Mark, also note that what Patrick has done is probably the best PETSc can
> do on GPUs, namely polynomial smoothers and block AIJ interpolators. I think
> that can all work with master right now, but you should ask him.
>

Thanks Matt,

At this point I am just answering questions ... but if you have smoothers
and interpolators (and a residual) that is all you need for the solve
phase, right? These folks have a linear Poisson solve so we can amortize
setup.


>
>    Matt
>
>
>> The fusion folks that I work with, and I assume other DOE offices, are
>> just looking at their codes, subroutine by subroutine, and having postdocs
>> look at GPUising them. We just need intelligent answers to their questions.
>> Even if we as sentient and passionate human being have opinions on the
>> approach that is implied by their questions, it is part of my job to just
>> give them a professional answer.
>>
>> I have enough now (thanks Jed and Lorena, et al!) to answer the AMGx
>> question sufficiently, and if you could give me a quick assessment of where
>> we are with hypre's GPU solver that would be great.
>>
>> Thanks,
>> Mark
>>
>>
>> On Thu, Jul 13, 2017 at 11:16 PM, Karl Rupp <rupp at iue.tuwien.ac.at>
>> wrote:
>>
>>> Hi Mark,
>>>
>>> I hear Hypre has support for GPUs in a May release. Any word on the
>>>> status of using it in PETSc?
>>>>
>>>
>>> as far as I know, it is currently not supported in PETSc. I'll have a
>>> look at it and see what needs to be done to enable it.
>>>
>>>
>>> And we discussed interfacing to AMGx, which is complicated (precluded?)
>>>> by not releasing source. Anything on the potential of interfacing to AMGx?
>>>> I think it would be great to make this available. It is on a lot of
>>>> checkboxes. I would love to be able to say, yea you can use it.
>>>>
>>>
>>> Lorena Barba's group actually interfaced PETSc to AMGx at some point
>>> (presented at GTC 2016 if I'm not mistaken). I'll reach out to them, maybe
>>> they have something to contribute.
>>>
>>> Best regards,
>>> Karli
>>>
>>
>>
>
>
> --
> What most experimenters take for granted before they begin their
> experiments is infinitely more interesting than any results to which their
> experiments lead.
> -- Norbert Wiener
>
> http://www.caam.rice.edu/~mk51/
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20170714/9fb74158/attachment.html>