Mailman 3 January 2015 - yt-users

parallel computation
by Yuxiang Qin 17 Jan '15

17 Jan '15

Dear YT-Users, I am a fresh man to YT-project. I got a few questions for parallel computation on yt. May I ask for help? I tried to do mpirun -np 4 myscript.py , but it always failed even when I used very simple script like #!/usr/bin/env python import yt yt.enable_parallelism() fname=('../../Smaug/NOZCOOL_LateRe_L010N0512/data/snapshot_103/snap_103.0.hdf5') ds = yt.load(fname) pz = yt.ProjectionPlot(ds, 'z', ('gas', 'temperature'),weight_field='density') pz.save() I am using the data from Dr. Alan Duffy. His data is similar to OWLS. So when I tried to run a similar script with sample data downloaded from yt website, it works. Does anyone know what’s going on here? Is anything wrong with my data format or script? How to check it? Warmest sincerely, -- Yuxiang Qin PhD Student School of Physics The University of Melbourne VIC, Australia, 3010 _______________________________________________ yt-users mailing list yt-users(a)lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

3 4

Re: [yt-users] cut_region: can I make an expression dimensionless?
by Stuart Levy 17 Jan '15

17 Jan '15

Oh, that sounds valuable. I ended up adding it at a low level- there is now a FilteredIOBrick, which takes a (list of) filter callbacks and takes under a second. But a yt-ish way to do it would be nicer. -------- Original message -------- From: Matthew Turk <matthewturk(a)gmail.com> Date: 01/16/2015 18:24 (GMT-06:00) To: Discussion of the yt analysis package <yt-users(a)lists.spacepope.org> Subject: Re: [yt-users] cut_region: can I make an expression dimensionless? Hi Stuart, Actually the more I look at this, the more I think it's probably quite easy to turn this into a data selector itself, avoiding the need for a cut region. Should be straightforward to add it to selection_routines and selection_data_containers and have it work out of the box. On Fri, Jan 16, 2015 at 6:20 PM, Matthew Turk <matthewturk(a)gmail.com> wrote: On Wed, Jan 14, 2015 at 11:03 AM, Stuart Levy <salevy(a)illinois.edu> wrote: Using one of Sam's past e-mails to answer my own question, it at least syntactically works to use the '.uq' unit-quantity attribute of a YT object to get something that has the units of the object, so e.g. 0.78 * obj['x'].uq is 0.78 in x's length units. So it actually seems to work to say: cr = ds.cut_region( dd, ["obj['z']-(obj['x']**2 + obj['y']**2)/(2.0*obj['z'].uq) < 0.78*obj['z'].uq"]) Neato! As it turns out, I'll look for another lower-level way to do this - with the above cut_region applied, 32GB of RAM isn't enough to process a single data field of about 512^3, and it takes several minutes of CPU time even to start. But it's good to know that the above is possible. Glad you kind of got it working ... this is kind of ridiculous that it takes so much, though. What I think could cut it dowen considerably is to reduce the numebr of sequential numpy operations; right now it's computing them and storing temporary arrays like crazy. Unfortunately I don't know how one might do this with the current setup of cut_region. Perhaps one way would be to make a derived field that is each component, and inside it do each operation inline, then apply the conditional there? On 1/14/15 10:26 AM, Stuart Levy wrote: Hello yt people, I'm hoping to write a cut_region expression that (a) depends on position and (b) does it in a nonlinear way - I'd like to do a sort of paraboloidal cut, like ds.cut_region(dd, "obj['z'] - (obj['x']**2 + obj['y']**2)/2.0 < 0.78") er, correction - cut_region()'s 2nd arg is a [] list of strings rather than a string, so I'd tried ds.cut_region(dd, ["obj['z'] - (obj['x']**2 + obj['y']**2)/2.0 < 0.78"]) But this runs afoul of the unit-checking - obj['z'] doesn't have the same units as obj['x']**2. Somehow I either need to cast all the obj[] terms to be dimensionless ("trust me, I promise it's right"), or else give dimensions of length to the constants. Should there be a way to do either one? _______________________________________________ yt-users mailing list yt-users(a)lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

1 0

velocity annotation dark matter
by Long Do Cao 17 Jan '15

17 Jan '15

Hi all yt-users, I'm running into a simple problem I think, but can't find the correct documentation. I want to make a slice plot of dark matter density with velocity vectors overlaid, using Gadget HDF5 data. I have uploaded my script here : http://paste.yt-project.org/show/5347/ The slice plot works fine. However, I got an error message for the velocity overlay: http://paste.yt-project.org/show/5349/ The bottom line is the following : YTFieldNotFound: Could not find field '('all', 'velocity_x')' in snapshot_000. The error is pretty obvious, I think YT is looking for the default field, whereas it should look for a deposit of the dark matter data. Somehow I should specify that yt should look for this deposit, but I don't know how ? Thank you by advance for your help, Long -------------------------------------------------------------- Long DO CAO (Office F617) Office phone: +81 6-6850-5482 Osaka University, Graduate School of Science Department of Earth and Space Science 1-1 Machikaneyama-cho Toyonaka, Osaka, 560-0043, Japan ---------------------------------------------------------------

2 2

Trying to use 16 CPUs via a python notebook
by Rick Sarmento 17 Jan '15

17 Jan '15

Hello, I’m pretty new to yt… I’m trying to speed things up significantly since I have rather large RAMSES datasets (running 3Mpc sims with 2^9^3 particles). I’m using STAMPEDE at TACC, https://vis.tacc.utexas.edu/# , to startup a notebook on their visualization nodes (each has 16 cores). I’m trying to use import yt yt.enable_parallelism() ds = yt.load("./output_00011/info_00011.txt”) and then # Looking through the "z" axis... Plot temp weighted by density p = yt.ProjectionPlot(ds, "z", "temperature", weight_field="density") p.annotate_title(z) p.show() but I only see one python process via top … and NO speedup. What do I need to do to take advantage of the 16 cores? Cheers! Rick Sarmento SESE Astronomy/Astrophysics Grad Student rsarment(a)asu.edu

2 1

RAMSES particles - star IDs are < 0
by Rick Sarmento 17 Jan '15

17 Jan '15

So I’m trying to “p.annotate_particles” with RAMSES star particles (over a temperature projection plot) …. But everything in RAMSES is a particle, so when I do p.annotate_particles I end up with a solid black plot… How do I tell annotate_particles to just use the particles with ID < 0 (these are the star particles)?? Thx! Cheers, Rick Sarmento earnric(a)gmail.com

4 3

cut_region: can I make an expression dimensionless?
by Stuart Levy 17 Jan '15

17 Jan '15

Hello yt people, I'm hoping to write a cut_region expression that (a) depends on position and (b) does it in a nonlinear way - I'd like to do a sort of paraboloidal cut, like ds.cut_region(dd, "obj['z'] - (obj['x']**2 + obj['y']**2)/2.0 < 0.78") But this runs afoul of the unit-checking - obj['z'] doesn't have the same units as obj['x']**2. Somehow I either need to cast all the obj[] terms to be dimensionless ("trust me, I promise it's right"), or else give dimensions of length to the constants. Should there be a way to do either one?

3 4

ANN: yt-3.1 released
by John ZuHone 15 Jan '15

15 Jan '15

The yt community is proud to announce the release of yt 3.1. yt (http://yt-project.org) is an open source, community-developed toolkit for analysis and visualization of volumetric data of all types, with a particular emphasis on astrophysical simulations and nuclear engineering simulations. This is a scheduled feature release. Highlighted changes in yt 3.1: Major changes: ++++++++++++++ * The RADMC-3D export analysis module has been updated. * Performance improvements for grid frontends. * Added a frontend for Dark Matter-only NMSU Art simulations. * The absorption spectrum generator has been updated. * The PerspectiveCamera has been updated and a new SphericalCamera has been added. * The unit system now supports unit equivalencies and has improved support for MKS units. * Data object selection can now be chained, allowing selecting based on multiple constraints. * Added the ability to manually override the simulation unit system. * The documentation has been reorganized and has seen substantial improvements. Minor or bugfix changes: ++++++++++++++++++++++++ * The Gadget InternalEnergy and StarFormationRate fields are now read in with the correct units. * Substantial improvements for the PPVCube analysis module and support for FITS datasets. * The center of a PlotWindow plot can now be set to the maximum or minimum of any field. * Projections are now performed using an explicit path length field for all coordinate systems. * Fix for the camera.draw_line function. * Minor fixes and improvements for yt plots. * Significant documentation reorganization and improvement. * Miscellaneous code cleanup. * yt now hooks up to the python logging infrastructure in a more standard fashion, avoiding issues with yt logging showing up with using other libraries. * Improvements for the yt-rockstar interface. * It is now possible to supply a default value for get_field_parameter. * A bug in the interpretation of the units of RAMSES simulations has been fixed. * Improvements and bugfixes for the halo analysis framework. * Fix issues with the default setting for the "center" field parameter. * yt can now be run in parallel on a subset of available processors using an MPI subcommunicator. * Fix for incorrect units when loading an Athena simulation as a time series. * Improved support for Enzo 3.0 simulations that have not produced any active particles. * Fix for periodic radius vector calculation. * Improvements for the Maestro and Castro frontends. * Clump finding is now supported for more generic types of data. * Fix unit consistency issue when mixing dimensionless unit symbols. * Improved memory footprint in the photon_simulator. * Large grids in Athena datasets produced by the join_vtk script can now be optionally split, improving parallel performance. * Slice plots now accept a “data_source" keyword argument. * Nearest neighbor distance field added. * Improvements for the ORION2 frontend. * Enzo 3.0 frontend can now read active particle attributes that are arrays of any shape. * Fixes for accessing deposit fields for FLASH data. * Added wrapper functions for numpy array manipulation functions. * Added support for packed HDF5 Enzo datasets. A more comprehensive list of the changes in this release, with links to the corresponding pull requests, can be found at http://yt-project.org/docs/3.1/reference/changelog.html. Standard Installation Methods ----------------------------- As with previous releases, you can install yt from source using one of the following methods. 1) From the install script (http://yt-project.org/#getyt): # Installation $ wget http://hg.yt-project.org/yt/raw/stable/doc/install_script.sh $ bash install_script.sh # Update $ yt update 2) From pip (source or binary wheel, see below for more details): # Installation $ pip install yt # Update $ pip install -U yt 3) From the Anaconda Python Distribution (https://store.continuum.io/cshop/anaconda/): # Installation $ conda install yt # Update $ conda update yt Note that it might take a day or two for the conda package to be updated. If you are on the “stable” branch, updating will bring you from yt 3.0.2 to 3.1, incorporating all changes since 3.0.2, whereas if you are on the “dev” or “yt” branch, only the changes since your last update should be incorporated. NEW: Installing Binary Packages via pip --------------------------------------- New to this release is the ability to install binary packages (“wheels”) using pip on Windows and Mac OS X (64-bit only for both). This has the advantage of not needing to install yt from source using a proper compiler setup, which has caused occasional problems on both of these platforms and prevented us from installing yt easily on other Python distributions. We have so far been able to install and run the binary distribution via pip on the following platforms and Python stacks: Windows x86_64: * Enthought Canopy Python (https://www.enthought.com/products/canopy/) * WinPython (http://winpython.sourceforge.net/) Mac OS X x86_64: * Enthought Canopy Python (https://www.enthought.com/products/canopy/) * Homebrew Python (http://brew.sh/) * Python.org Python * Mac OS X’s system Python * MacPorts Python (https://www.macports.org/) This is somewhat experimental, so other distributions may work (or not), please submit bug reports or successes to the mailing list or to the Bitbucket issues page (http://bitbucket.org/yt_analysis/yt/issues). All distributions must be Python v. 2.7. The requirements for installing yt via this method are the same as from source: * NumPy * h5py * HDF5 * SymPy * Matplotlib * IPython (not required, but strongly recommended) To install a new version of yt on one of these platforms, simply do $ pip install yt and you should get the binary distribution automatically. Also, if your python installation is system-wide (e.g., the Mac system Python) you might need to run pip with administrator privileges. For more information, including more installation instructions, links to community resources, and information on contributing to yt’s development, please see the yt homepage at http://yt-project.org and the documentation for yt-3.1 at http://yt-project.org/docs/3.1. yt is the product of a large community of developers and users and we are extraordinarily grateful for and proud of their contributions. Please forward this announcement on to any interested parties. As always, if you have any questions, concerns, or run into any trouble updating please don't hesitate to send a message to the mailing list or stop by our IRC channel. Thank you, The yt development team

1 0

Axes ordering for SlicePlot
by Stuart Mumford 10 Jan '15

10 Jan '15

Hello, I am plotting a 2D cut of my domain using SlicePlot, is there a way to flip the axes ordering for normal='y' so the z axis is on the vertical? Example: http://i.imgur.com/QULeYt2.png Thanks Stuart

2 1

Place plot on a specified matplotlib axis
by David Ketcheson 09 Jan '15

09 Jan '15

I would like to insert a yt-generated plot in a particular matplotlib axes object that I created. In matplotlib, I can do this via fig2 = plt.figure() ax2 = fig2.add_axes([0.15, 0.1, 0.7, 0.3]) ax2.plot(x,y) which will place the plot on the axes ax2. Is there any way to do this in yt? ________________________________ This message and its contents including attachments are intended solely for the original recipient. If you are not the intended recipient or have received this message in error, please notify me immediately and delete this message from your computer system. Any unauthorized use or distribution is prohibited. Please consider the environment before printing this email.

3 2

Logging of output
by Max Katz 09 Jan '15

09 Jan '15

Hi all, A glance at the source code (utilities/logger.py) indicates to me that yt is using the Python logging package for output information, and that depending on the choice of the option stdoutStreamLogging, this can be directed to either stdout or stderr. Since there's a lot of information piped to stdout whenever a dataset is opened, and this can tend to overwhelm any other information that is being displayed to screen it would be nice if the logging output could be redirected to a file. This can be done when setting the basicConfig for the logger <https://docs.python.org/2/howto/logging.html> when setting it up, but that page indicates that this configuration cannot be overridden after it has initially been set. Is there a way to include an option in yt for the user to direct the logging output to a file of their choice? Max Katz Department of Physics and Astronomy Vice President, Graduate Student Organization Stony Brook University http://astro.sunysb.edu/mkatz/

3 2