On Mon, Jul 24, 2017 at 12:42 PM, Karl Rupp <[email protected]> wrote:

> Hi folks,
>
> alright, let me summarize and organize things. If you're on the list,
> please provide or update information on 'what left' and 'expected
> completion':
>
> * SF support for MPIUNI
>    - Who: Jed
>    - What left: Implementation
>    - Expected completion: ?
>
> * Hashtable support for matrix assembly
>    - Who: Jed
>    - What left: ?
>    - Expected completion: ?
>
> * New GLVis based viewer for DMDAs, FEM an FVM
>    - Who: Stefano
>    - What left: Test & merge
>    - Expected completion: in a few days
>
> * Tools for particle-in-cell methods
>    - Who: Dave
>    - What left: ?
>    - Expected completion: ?
>
> * AIJMKL submatrix type
>    - Who: Richard
>    - What left: Cleanup, test & merge
>    - Expected completion: a few days
>
> * Manual chapter: Performance Hints
>    - Who: Karli
>    - What left: More writing
>    - Expected completion: this week
>

Karl: 1) thanks for the summary, and 2) Please let me know when your think
your section of the manual is in a state in which I can look at it and make
some contributions. One thing I'd like to write about (if you haven't
already covered it) is some basic info about controlling MPI process
placement/pinning and the (sometimes surprisingly large) effects it can
have on performance. This is getting a lot more complicated as systems add
more NUMA domains and hardware threads. When I was at Intel I encountered a
ton of performance problems that were mostly due to bad process placement
(which, fortunately, meant they were actually easy to fix!).

--Richard


> I thought about using the issue tracker for this, but without support
> for proper tagging or milestones the issue tracker on Bitbucket is of
> very limited use.
>
> Best regards,
> Karli
>
>
> On 07/23/2017 05:08 PM, Barry Smith wrote:
> >
> >     Anything stopping us from making a PETSc release?
> >
> >     Barry
> >
>

Reply via email to