I recommend using openmpi. I have been able to build openmpi on multiple platforms and then build mpi4py with it without any customization. As Matt has said, though, you won't see any benefit to using parallel until your simulations are at least 256^3 cells or more.<br>
<br><div class="gmail_quote">On Thu, Feb 12, 2009 at 8:16 PM, Matthew Turk <span dir="ltr"><<a href="mailto:matthewturk@gmail.com">matthewturk@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Hi again,<br>
<br>
I just realized that I should say a couple important caveats --<br>
<br>
1. We haven't released 'yt-trunk' as 1.5 yet because it's not quite<br>
done or stable. It's going well, and many people use it for<br>
production-quality work, but it's not really stamped-and-completed.<br>
2. I should *also* note that you won't really get a lot out of<br>
parallel yt unless you have relatively large datasets or relatively<br>
large amounts of computation on each cell while creating a derived<br>
field. It might end up being a bit more work than you're looking for,<br>
if you just want to get some plots out quickly.<br>
<br>
-Matt<br>
<div><div></div><div class="Wj3C7c"><br>
On Thu, Feb 12, 2009 at 7:12 PM, Matthew Turk <<a href="mailto:matthewturk@gmail.com">matthewturk@gmail.com</a>> wrote:<br>
> Hi!<br>
><br>
> yt-trunk is now parallelized. Not all tasks work in parallel, but<br>
> projections, profiles (if done in 'lazy' mode) and halo finding (if<br>
> you use the SS_HopOutput module) are now parallelized. Slices are<br>
> almost done, and the new covering grid will be. It's not documented,<br>
> but those tasks should all run in parallel. We will be rolling out a<br>
> 1.5 release relatively soon, likely shortly after I defend my thesis<br>
> in April, that will have documentation and so forth.<br>
><br>
> I'm surprised you can't compile against the mpich libraries in a<br>
> shared fashion. Unfortunately, I'm not an expert on MPI<br>
> implementations, so I can't quite help out there. In my personal<br>
> experience, using OpenMPI, I have needed to except when running on<br>
> some form of linux without a loader -- the previous discussion about<br>
> this was related to Kraken, which runs a Cray-specific form of linux<br>
> called "Compute Node Linux." I don't actually know offhand (anybody<br>
> else?) of any non-Cray machines at supercomputing out there require<br>
> static linking as opposed to a standard installation of Python. (I'm<br>
> sure they do, I just don't know of them!)<br>
><br>
> As for the second part, usually when instantiating you have to run the<br>
> executable via mpirun. (On other MPI implementations, this could be<br>
> something different.) One option for this -- if you're running off<br>
> trunk -- would be to do something like:<br>
><br>
> mpirun -np 4 python my_script.py --parallel<br>
><br>
> where the file my_script.py has something like:<br>
><br>
> --<br>
> from yt.mods import *<br>
> pf = EnzoStaticOutput("my_output")<br>
> pc = PlotCollection(pf, center=[0.5,0.5,0.5])<br>
> pc.add_projection("Density",0)<br>
> pc.save("hi_there")<br>
> --<br>
><br>
> The projection would be executed in parallel, in this case. (There is<br>
> a command line interface called 'yt' that also works in parallel, but<br>
> it's still a bit in flux.) You can't just run "python" because of the<br>
> way the stdin and stdout streams work; you have to supply a script, so<br>
> that it can proceed without input from the user. (IPython's parallel<br>
> fanciness notwithstanding, which we do not use in yt.)<br>
><br>
> But, keep in mind, running "mpirun -np 4" by itself, wihtout setting<br>
> up a means of distributing tasks (usually via a tasklist) will run<br>
> them all on the current machine. I am, unfortunately, not really<br>
> qualified to speak to setting up MPI implementations. But please do<br>
> let us know if you have problems with the yt aspects of this!<br>
><br>
> -Matt<br>
><br>
> On Thu, Feb 12, 2009 at 6:59 PM, rsoares <<a href="mailto:dlleuz@xmission.com">dlleuz@xmission.com</a>> wrote:<br>
>> Hi,<br>
>><br>
>> I'm trying to run mpi4py on my 4 machines, but I need a parallelized version<br>
>> of Python. Tried to compile one with Python 2.5 and mpich2 but mpich2 won't<br>
>> let me built dynamic /shares libraries which it needs. Trying with the<br>
>> static ones involves alot of headers errors from both.<br>
>> Is yt-trunk capable of doing python in parallel?<br>
>><br>
>> Without parallel-python, I mpdboot -n 4 then<br>
>><br>
>> python<br>
>>>>>import MPI<br>
>>>>> rank, size = MPI.COMM_WORLD.rank, MPI.COMM_WORLD.size<br>
>>>>> print 'Hello World! I am process', rank, 'of', size<br>
>> Hello World! I am process 0 of 1<br>
>>>>><br>
>><br>
>> not 4 processes, and mpirun -np 4 python just hangs. mpi4py installed on<br>
>> all 4 nodes.<br>
>><br>
>> Thanks.<br>
>><br>
>> R.Soares<br>
>> _______________________________________________<br>
>> yt-users mailing list<br>
>> <a href="mailto:yt-users@lists.spacepope.org">yt-users@lists.spacepope.org</a><br>
>> <a href="http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org" target="_blank">http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org</a><br>
>><br>
><br>
_______________________________________________<br>
yt-users mailing list<br>
<a href="mailto:yt-users@lists.spacepope.org">yt-users@lists.spacepope.org</a><br>
<a href="http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org" target="_blank">http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org</a><br>
</div></div></blockquote></div><br>