[yt-users] parallel issue

[nick moeckel]

ah, those errors in [1] look very familiar. Switching to yt.mods seems to
work, thanks for the help!

Nick


On Mon, Nov 4, 2013 at 9:44 AM, Nathan Goldbaum <nathan12343 at gmail.com>wrote:

> Hi Nick,
>
> Are you sure about that?  The traceback you pasted was from a script that
> failed on "from yt.pmods import *".
>
> In any case, there are issues with pmods and matplotlib 1.3.0 that, as far
> as I am aware, have not been resolved [1].
>
> pmods is an implementation of a fast parallel importer to avoid slowdowns
> that occur on massively parallel imports of yt's namespace.  If you're
> running at scale on a supercomputer, pmods might be useful, but it is not
> necessary to get yt's parallelism working.  A script with pmods replaced
> with mods will still work correctly in parallel.
>
> There is an open issue to update the documentation to explain what pmods
> does and to put less emphasis on pmods in the examples [2].
>
> -Nathan
>
> [1]
> https://bitbucket.org/yt_analysis/yt/issue/637/matplotlib-130-doesnt-seem-compatible-with
> [2]
> https://bitbucket.org/yt_analysis/yt/issue/672/update-docs-that-pmods-not-required-for-yt
>
>
> On Mon, Nov 4, 2013 at 12:28 AM, nick moeckel <nickolas1 at gmail.com> wrote:
>
>> Hi Geoffrey,
>>
>> yt is up to date and working, I've been using it serially with no problem
>> on these machines. I'm able to import yt.pmods with no problems.
>>
>> Nick
>>
>>
>> On Mon, Nov 4, 2013 at 8:40 AM, Geoffrey So <gsiisg at gmail.com> wrote:
>>
>>> Since it failed on line 1, where you import yt.pmods, that's not an
>>> explicit mpi4py problem I think.
>>>
>>> Have you made sure yt is installed correctly?  Are you able to
>>> >python
>>> from yt.mods import *
>>>
>>> and if that works, try replacing yt.pmods with yt.mods
>>>
>>> Hope that helps!
>>>
>>> From
>>> G.S.
>>>
>>>
>>> On Sun, Nov 3, 2013 at 11:17 PM, nick moeckel <nickolas1 at gmail.com>wrote:
>>>
>>>> Hi everyone,
>>>>
>>>> I'm trying to run some parallel analysis scripts and I'm running into
>>>> problems. I'm using yt-3.0 on Ramses data; here's a simple example that
>>>> doesn't work:
>>>>
>>>> from yt.pmods import *
>>>> ds = load('output_00001/info_00001.txt',
>>>> fields=['Density','x-velocity','y-velocity','z-velocity','Pressure'])
>>>> proj = ProjectionPlot(ds, 0, 'Density')
>>>> proj.save()
>>>>
>>>> the errors this generates:
>>>> http://paste.yt-project.org/show/SfT87oaHcQpYMUnBi2QM/
>>>>
>>>> When I run this on my personal computers with mpi4py installed with
>>>> pip, this will work with no problems. On the HPC system I'm using I'm
>>>> unable to install mpi4py with pip, (due [I think] to the archaic openssl
>>>> version) but I've built and installed it from source.
>>>>
>>>> Before I tear any more of my hair out over this, I'm hoping that
>>>> someone can tell me if this looks familiar, or if it's definitely a problem
>>>> with mpi4py and yt playing nice, or anything else that might save me a few
>>>> days of my life.
>>>>
>>>> Thanks for any help!
>>>> Nick
>>>>
>>>> _______________________________________________
>>>> yt-users mailing list
>>>> yt-users at lists.spacepope.org
>>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>>>
>>>>
>>>
>>> _______________________________________________
>>> yt-users mailing list
>>> yt-users at lists.spacepope.org
>>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>>
>>>
>>
>> _______________________________________________
>> yt-users mailing list
>> yt-users at lists.spacepope.org
>> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>>
>>
>
> _______________________________________________
> yt-users mailing list
> yt-users at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.spacepope.org/pipermail/yt-users-spacepope.org/attachments/20131104/a5edd3af/attachment.html>