[Wien] Memory leak in impi -- please cross-report

[Laurence Marks]

N.B., you can also add to https://software.intel.com/en-us/comment/1930019
if you have seen the memory leak...

On Wed, Feb 13, 2019 at 7:01 AM Laurence Marks <L-marks at northwestern.edu>
wrote:

> Some months ago I reported a severe memory leak in lapw1_mpi with Intel
> impi. The leak is present for certain in all 2019 versions, but goes away
> if one regresses the version (I use 2015). I know of one other Wien2k user
> who has found it, and I recently found it on an AMD test node.
>
>  I reported it to Intel on  11/26/2018 and it is request 03837257. Beyond
> a software engineer at Intel verifying it and indicating that it is
> probably in impi (which I already knew) it is going nowhere. Perhaps some
> of you can cross-report it, maybe this will help.
>
> I strongly suspect that the bug is linked to PDSYGST in seclr4.F. Based
> upon my tests this is (with elpa) the road block, as it does not scale with
> the number of cores used (i.e. time taken is essentially independent of
> cores) for larger problems (e.g. 50K matrix size)
>
> --
> Professor Laurence Marks
> "Research is to see what everybody else has seen, and to think what nobody
> else has thought", Albert Szent-Gyorgi
> www.numis.northwestern.edu ; Corrosion in 4D:
> MURI4D.numis.northwestern.edu
> Partner of the CFW 100% program for gender equity, www.cfw.org/100-percent
> Co-Editor, Acta Cryst A
>


-- 
Professor Laurence Marks
"Research is to see what everybody else has seen, and to think what nobody
else has thought", Albert Szent-Gyorgi
www.numis.northwestern.edu ; Corrosion in 4D: MURI4D.numis.northwestern.edu
Partner of the CFW 100% program for gender equity, www.cfw.org/100-percent
Co-Editor, Acta Cryst A
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20190213/9f0f3e0a/attachment.html>