yt-dev

Download

yt-dev@python.org

February 2011

11 participants
12 discussions

HEALpix volume rendering
by Matthew Turk Feb. 10, 2011

Feb. 10, 2011

Hi all, Just wanted to let you know that I've put in experimental support for volume rendering using HEALpix to define the direction of vectors, all of which will then start at the camera location and then move radially outwards, to give a set of rays that cover the entire sky with equal area. ( http://healpix.jpl.nasa.gov/ ) This is not adaptive in any way; these rays will not split along their trajectories, which is broadly problematic. Additionally, the bricks are not sorted yet, so currently only projection transfer functions (column density) are supported. I'll add support for ordering the grids so that more complicated things that require inputs from previous bricks can be accomplished, but for now it's only useful for column density and averaging operations. Anyway, if you're interested in playing with this, I've put up a script here: http://paste.enzotools.org/show/1498/ And I've extracted portions of the healpy module that are necessary to plot a Molleweide projection, which can be found here: http://yt.enzotools.org/files/mv.txt note that once you download, you have to change mv.txt to mv.py. Attached is an image from one of the test simulations we used at the Enzo workshop, where I've tossed out nside=64 (12*64*64 vectors) out to a radius of 0.1 of the box. I then plotted the column density. This is the output from the linked script, above. This functionality is also able to do angular-averages of values, which is kind of cool (and what I am applying it to.) My plan is to add correct ordering of grids, create a Molleweide projection that is small and works like other yt plots, and then to make it easier to generate and plot. If anybody wants to help out, I'd be appreciative -- particularly with the generation of projections and plotting of them. -Matt

4 3

errors in SpecificAngularMomentum
by John Wise Feb. 9, 2011

Feb. 9, 2011

Hi all, I was trying to calculate the spin parameter for a halo with the following simple script, center = [0.495013, 0.494046, 0.495160] radius = 100 # pc pf = load("DD0019/output_0019") sp = pf.h.sphere(center, radius/pf["pc"]) gas_spin = sp.quantities["BaryonSpinParameter"]() and it doesn't work in parallel. It gives the following traceback. --- Traceback (most recent call last): File "spin.py", line 9, in <module> gas_spin = sp.quantities["BaryonSpinParameter"]() File "/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages/yt-2.1dev-py2.6-macosx-10.6-intel.egg/yt/data_objects/derived_quantities.py", line 85, in __call__ self.func(e, *args, **kwargs) File "/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages/yt-2.1dev-py2.6-macosx-10.6-intel.egg/yt/data_objects/derived_quantities.py", line 237, in _BaryonSpinParameter am = data["SpecificAngularMomentum"]*data["CellMassMsun"] ValueError: shape mismatch: objects cannot be broadcast to a single shape Traceback (most recent call last): File "spin.py", line 9, in <module> gas_spin = sp.quantities["BaryonSpinParameter"]() File "/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages/yt-2.1dev-py2.6-macosx-10.6-intel.egg/yt/data_objects/derived_quantities.py", line 85, in __call__ self.func(e, *args, **kwargs) File "/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages/yt-2.1dev-py2.6-macosx-10.6-intel.egg/yt/data_objects/derived_quantities.py", line 237, in _BaryonSpinParameter am = data["SpecificAngularMomentum"]*data["CellMassMsun"] ValueError: shape mismatch: objects cannot be broadcast to a single shape --- This happens on multiple systems. I looked into this further, and I think the shapes of the angular momentum data are not being preserved in the MPI calls. Also the shapes appear wrong in a serial calculation. In serial, everything looks fine until data["SpecificAngularMomentum"] is returned. Here are the shapes of the arrays. v_vec: (3,16,16,16) cross product of r_vec and v_vec: (3,16,16,16) and the returned array from _SpecificAngularMomentum has a shape of (3,0)! Then when I try to run it with 2 processors, the shapes change to v_vec: (3,4096) cross: (3,4096) _SpecificAngularMomentum result: (12288,) Can someone replicate this error? Any ideas? Thanks, John

2 2