Mailman 3 January 2012 - yt-dev

Call for testing: SHA512 in the install_script instead of md5 on the server
by Matthew Turk 01 Feb '12

01 Feb '12

Hi all, Currently to verify file integrity with the install script we download sidecar md5 file from yt-project.org. These aren't versioned, and md5 itself is largely deprecated for this. I've created a new sha512-using version of the install script, which uses sha512 hashes stored *in the install script* to verify file integrity. This way any time these change we will be notified of them. I'd like it if a few people could test this -- I have -- on pristine systems. It changes both how the files are downloaded and it now uses sha512sum, which should be available on most systems (according to Kacper :). You can do this by: wget https://bitbucket.org/MatthewTurk/yt/raw/367ea3bfff2e/doc/install_script.sh and then running it, but maybe supplying an alternate directory rather than the default. Thanks for any feedback. -Matt

6 13

yt workshop 2012: A success!
by Matthew Turk 30 Jan '12

30 Jan '12

Hi all, The yt workshop last week in Chicago ( http://yt-project.org/workshop2012/ ) was an enormous success. On behalf of the organizing and technical committees, I'd like to specifically thank the FLASH Center, particularly Don Lamb, Mila Kuntu, Carrie Eder, for their hospitality; the venue was outstanding and their hospitality touching. Additionally, we're very grateful to the Adler Planetarium's Doug Roberts and Mark SubbaRao for hosting us on Wednesday evening -- seeing the planetarium show as well as volume renderings made by yt users up on the dome was so much fun. The yt workshop was supported by NSF Grant 1214147. Thanks to everyone who attended -- your energy and excitement helped make it a success. Thanks also to the organizing and technical committees: Britton Smith, John ZuHone, Brian O'Shea, Jeff Oishi, Stephen Skory, Sam Skillman, and Cameron Hummels. All talks have been recorded, and you can clone a unified repository of talk slides and worked examples: hg clone https://bitbucket.org/yt_analysis/workshop2012/ A few photos have been put up online, too: http://goo.gl/g02uP As I am able to edit and upload talks, they'll appear on the yt youtube channel as well as on the yt homepage: http://www.youtube.com/ytanalysis Thanks again, and wow, what a week! Matt

1 0

Unit conversions in frontends
by Casey W. Stark 30 Jan '12

30 Jan '12

Hi all. I'm a bit confused about the unit mechanics in yt right now. I noticed there's a convert_function for fields, but there are also the units, time_units, and conversion_factors attributes for StaticOutput. For the Nyx frontend, I have added convert_function's for each field to convert the cosmological units into CGS. This works in my fork now and everything is in cgs and plots fine. Now I know I need to update the _set_units method of NyxStaticOutput, but I'm not sure what to do with those attributes and how they are used later. Best, Casey

2 3

Fwd: [mpi4py] [ANN] mpi4py release 1.3
by Matthew Turk 20 Jan '12

20 Jan '12

Hi all, New mpi4py -- pretty cool stuff. The lowercase irecv means non-blocking receive of a Python object, not just a buffer. -Matt ---------- Forwarded message ---------- From: Lisandro Dalcin <dalcinl(a)gmail.com> Date: Fri, Jan 20, 2012 at 2:40 PM Subject: [mpi4py] [ANN] mpi4py release 1.3 To: mpi4py(a)googlegroups.com Release 1.3 [2012-01-20] ======================== * Now ``Comm.recv()`` accept a buffer to receive the message. * Add ``Comm.irecv()`` and ``Request.{wait|test}[any|all]()``. * Add ``Intracomm.Spawn_multiple()``. * Better buffer handling for PEP 3118 and legacy buffer interfaces. * Add support for attribute attribute caching on communicators, datatypes and windows. * Install MPI-enabled Python interpreter as ``<path>/mpi4py/bin/python-mpi``. * Windows: Support for building with Open MPI. -- Lisandro Dalcin --------------- CIMEC (INTEC/CONICET-UNL) Predio CONICET-Santa Fe Colectora RN 168 Km 472, Paraje El Pozo 3000 Santa Fe, Argentina Tel: +54-342-4511594 (ext 1011) Tel/Fax: +54-342-4511169 -- You received this message because you are subscribed to the Google Groups "mpi4py" group. To post to this group, send email to mpi4py(a)googlegroups.com. To unsubscribe from this group, send email to mpi4py+unsubscribe(a)googlegroups.com. For more options, visit this group at http://groups.google.com/group/mpi4py?hl=en.

2 1

Ghost zone pull request
by Matthew Turk 17 Jan '12

17 Jan '12

Hi all, I put up a PR for much improved ghost zone filling (particularly for FLASH data!) that I'd appreciate a set of eyes on: https://bitbucket.org/yt_analysis/yt/pull-request/54/improve-ghost-zone-gen… Sam tested and sees that for some particular hierarchies there's a slight slowdown as a result, but I think the improvement where it applies will be worth that. Thanks, Matt

1 0

Fwd: [mpi4py] Fwd: [Numpy-discussion] Improving Python+MPI import performance
by Matthew Turk 16 Jan '12

16 Jan '12

For a long time we have been running into this very problem. I think it would be appropriate to utilize this code on Kraken, Ranger, etc. My implementation suggestion would be to put this in the new startup_tasks, where we determine parallelism. As noted in the docstring it will have to be modified to use mpi4py. Britton or Stephen, this sounds like it's directly up your alley as you run on Kraken the most often. Would one of you be willing to test it out? My feeling is that we could simply suggest that on these systems we use this idiom at the top of scripts (where we assume we distribute this script with yt): from yt.mpi_importer import mpi_import with mpi_import(): from yt.mods import * I think it should recursively watch all the imports. An alternate option would be to insert some of its logic into yt.mods, or even have a second mods file that handles it seamlessly, like: from yt.pmods import * Ideas? -Matt ---------- Forwarded message ---------- From: Dag Sverre Seljebotn <d.s.seljebotn(a)astro.uio.no> Date: Fri, Jan 13, 2012 at 3:51 AM Subject: [mpi4py] Fwd: [Numpy-discussion] Improving Python+MPI import performance To: mpi4py(a)googlegroups.com Cc: Chris Kees <cekees(a)gmail.com> This looks very interesting, Dag -------- Original Message -------- Subject: [Numpy-discussion] Improving Python+MPI import performance Date: Thu, 12 Jan 2012 17:13:41 -0800 From: Asher Langton <langton2(a)llnl.gov> Reply-To: Discussion of Numerical Python <numpy-discussion(a)scipy.org> To: numpy-discussion(a)scipy.org Hi all, (I originally posted this to the BayPIGgies list, where Fernando Perez suggested I send it to the NumPy list as well. My apologies if you're receiving this email twice.) I work on a Python/C++ scientific code that runs as a number of independent Python processes communicating via MPI. Unfortunately, as some of you may have experienced, module importing does not scale well in Python/MPI applications. For 32k processes on BlueGene/P, importing 100 trivial C-extension modules takes 5.5 hours, compared to 35 minutes for all other interpreter loading and initialization. We developed a simple pure-Python module (based on knee.py, a hierarchical import example) that cuts the import time from 5.5 hours to 6 minutes. The code is available here: https://github.com/langton/MPI_Import Usage, implementation details, and limitations are described in a docstring at the beginning of the file (just after the mandatory legalese). I've talked with a few people who've faced the same problem and heard about a variety of approaches, which range from putting all necessary files in one directory to hacking the interpreter itself so it distributes the module-loading over MPI. Last summer, I had a student intern try a few of these approaches. It turned out that the problem wasn't so much the simultaneous module loads, but rather the huge number of failed open() calls (ENOENT) as the interpreter tries to find the module files. In the MPI_Import module, we have rank 0 perform the module lookups and then broadcast the locations to the rest of the processes. For our real-world scientific applications written in Python and C++, this has meant that we can start a problem and actually make computational progress before the batch allocation ends. If you try out the code, I'd appreciate any feedback you have: performance results, bugfixes/feature-additions, or alternate approaches to solving this problem. Thanks! -Asher _______________________________________________ NumPy-Discussion mailing list NumPy-Discussion(a)scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion -- You received this message because you are subscribed to the Google Groups "mpi4py" group. To post to this group, send email to mpi4py(a)googlegroups.com. To unsubscribe from this group, send email to mpi4py+unsubscribe(a)googlegroups.com. For more options, visit this group at http://groups.google.com/group/mpi4py?hl=en.

4 15

accumulation in profiles
by david collins 13 Jan '12

13 Jan '12

Hi-- Is there a way to get a profile to accumulate from [bin: high bin], instead of [low bin: bin]? I poked through the source and I only see "accumulation = True, False", but not "\pm 1, 0". Thanks, d. -- Sent from my computer.

2 2

gdf fields
by Sam Skillman 04 Jan '12

04 Jan '12

Hi all, I'm running into an issue with translation dictionaries for the GDF frontend. On the tip, if I load up a gdf dataset, and do: pf.h.find_max('Density') I get: http://paste.yt-project.org/show/2008/ If I instead do pf.h.find_max('density'), I get what looks like the same error. If I replace line 56 in yt/frontends/gdf/fields.py 55 KnownGDFFields = FieldInfoContainer() 56 add_gdf_field = KnownGDFFields.add_field with 57 add_gdf_field = GDFFieldInfo.add_field It works fine. Alternatively, if I replace 96 for f,v in log_translation_dict.items(): 97 add_field(f, function=TranslationFunc(v), take_log=True) 98 99 for f,v in translation_dict.items(): 100 add_field(f, function=TranslationFunc(v), take_log=False) with 102 def _generate_translation(mine, theirs, take_log=False): 103 add_field(theirs, function=lambda a, b: b[mine], take_log=take_log) 104 105 106 for f,v in log_translation_dict.items(): 107 if v not in GDFFieldInfo: 108 add_field(v, function=lambda a,b: None, take_log=True, 109 validators = [ValidateDataField(v)]) 110 #print "Setting up translator from %s to %s" % (v, f) 111 _generate_translation(v, f, take_log=True) 112 113 114 for f,v in translation_dict.items(): 115 if v not in GDFFieldInfo: 116 add_field(v, function=lambda a,b: None, take_log=False, 117 validators = [ValidateDataField(v)]) 118 #print "Setting up translator from %s to %s" % (v, f) 119 _generate_translation(v, f, take_log=False) as is done in the orion reader, it works fine. Anyways, if someone sees what is going on here, let me know. I'll also be on IRC for a bit longer tonight and all tomorrow. Thanks, Sam

2 5