[FLASH-USERS] too many refinement iterations
Slavin, Jonathan
jslavin at cfa.harvard.edu
Wed Sep 25 08:30:20 EDT 2019
Hi Klaus,
Thanks for clearing that up. Still that's only 258 MB vs. total memory of
32 GB.
In fact I just redid the run with maxblocks set to 8000 on 10 processors
and it went for much longer. It did eventually fail with "Too many blocks!
Increase MAXBLOCKS or use more processors."
For me the takeaway is that when getting close to running out of blocks
FLASH will sometimes get into this mode where it iterates more and more
when refining. It's a symptom of running out of blocks, not running out of
memory (at least in my case).
Jon
On Tue, Sep 24, 2019 at 5:04 PM Klaus Weide <klaus at flash.uchicago.edu>
wrote:
> On Tue, 24 Sep 2019, Slavin, Jonathan wrote:
>
> > It looks like the total blocks is approaching the maximum allocated via
> > maxblocks, but I don't think the used memory is approaching the available
> > memory. My machine has 32 GB of memory. If I'm understanding the
> > calculation in the user's guide correctly the total memory requirement is
> > 14 (NUNK_VARS) * ( 8 (nguard) + 16 (nxb)) * ( 8 + 16 (nyb))* 4000
> > (maxblocks) = 32 MB
> > is that right?
>
> You forgot a factor of 8 (bytes per real number).
>
> Klaus
>
--
Jonathan D. Slavin
Astrophysicist - High Energy Astrophysics Division
Center for Astrophysics | Harvard & Smithsonian
Office: (617) 496-7981 | Cell: (781) 363-0035
60 Garden Street | MS 83 | Cambridge, MA 02138
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://flash.rochester.edu/pipermail/flash-users/attachments/20190925/d07dd69a/attachment-0001.htm>
More information about the flash-users
mailing list