Mailman 3 May 2012 - yt-users

A basic question how yt treat the intersect cell.
by Jun-Hwan Choi 08 May '12

08 May '12

Hi all, I have one conceptual question. When you make a profile, what happen one grid cell (finest cell) is bigger than one bin? Does yt interpolate the physical value (mass, density, velocity ...) and assign the appropriate value to the bins which the cell intersects? The very similar question is what happen one grid cell (finest) intersects with object boundary. For example, I define sphere and try to measure total mass of sphere. Does yt interpolate the mass of cell that intersects with the object's boundary and compute the partial mass of cell to assign to total mass of the sphere? If yt interpolate the cell, does yt use a particular weight? This issue may be important, when one need to compute fine physical quantities. Thanks in advance, Junhwan -- -------------------------------------------------------------- Jun-Hwan Choi, Ph.D. Department of Physics and Astronomy, University of Kentucky Tel: (859) 897-6737 Fax: (859) 323-2846 Email: jhchoi(a)pa.uky.edu URL: http://www.pa.uky.edu/~jhchoi --------------------------------------------------------------

4 5

Off-axis projections using Homogenized Volumes and a weight field
by Cameron Hummels 07 May '12

07 May '12

Hello everyone, I'm currently trying to generate an off-axis projection using a homogenized volume as an input, instead of the default kd-tree method. I have no problems using this method, until I try to add a weight field to the projection, then it simply returns "Inf" values for each pixel. I checked, and the exact same source using the kd-tree method works with no problems (it's just slower because I'm running this on a small subset of the overall dataset). Do I need to do something special when I create the homogenized volume to take into account the additional weighting field? I didn't see a function parameter for a weight in the constructor for HomogenizedVolume, and looking at the source doesn't provide any obvious clues, but perhaps I just need to code it in. Suggestions? Cameron

3 2

Interpreting parallel HOP catalog results
by Matthew Turk 04 May '12

04 May '12

Hi all, Has anybody seen something like this before? This is output from a parallelHF catalog: http://paste.yt-project.org/show/2340/ What I think is odd is that the first set of spatial coordinates are the same for all the halos. Is that desired behavior? Thanks, Matt

3 7

python timing
by Eric Hallman 03 May '12

03 May '12

Is there a standard or best methodology for tracking python code performance? I see in the intertubes that one can use the time module, or timeit or other things. How about memory usage? Or is this typically handled by debugging tools? What I really want to do is isolate logjams in python code by function call etc. in a fairly non-intrusive way. Thanks for any advice. Eric -- Eric Hallman Tech-X Corporation hallman(a)txcorp.com 5621 Arapahoe Ave, Suite A Phone: (720) 254-5833 Boulder, CO 80303 Fax: (303) 448-7756 --

3 3

Projection Performance
by Richard P Wagner 03 May '12

03 May '12

Hi, I wanted to build a sequence of projections using various color maps along each axis. The data set I'm using the z = 0 one from the L7 simulation some of you are familiar with. Here's the paste of my current script: http://paste.yt-project.org/show/2335/ (The early sys.exit is deliberate.) Does anyone have an estimate of the time doing this projection serially should take? After two hours it was still going without having produced the first image. The same plots done on the z = 2.75 data took about 15 minutes (this data has about 1/5 the number of grids, though). I will also gladly take advice on better methods for the projections, or if the benefit to doing this in parallel is worth it. Thanks, Rick

4 6

orientation of renderings and write_bitmap
by Sam Skillman 02 May '12

02 May '12

Hi all, There have been a few changes recently dealing with the orientation of the image that come out of the volume rendering. In https://bitbucket.org/yt_analysis/yt/pull-request/144/multi-index-widths-in…, the east and north vector were (rightly) swapped such that when specifying widths of the image it is done in (x,y,depth) instead of (y,x,depth). This has the side effect of transposing the images that are returned. However, this had been counteracted by the way write_bitmap takes the image and outputs it to png. So we were somewhat having two bugs counteract each other such that if you had requested north_vector to point in a particular direction, that showed up as "up". Once these changes went in, the images then showed up transposed and "up" became "right" (or maybe left). So, in a poorly-thought out action, I decided to just transpose the image in write_bitmap. This is a fairly ugly fix because now write_bitmap will always transpose any image that you throw it, so any other place you might use it in scripts might give strange behavior. We are aiming to fix this to work in a much more desirable fashion by the time we merge in the volume rendering refactor work that has been going on, but if anyone has any ideas on how we could handle it, it'd be great to hear. For now just be warned if you notice strange behavior with calls to write_bitmap. Best, Sam

1 0

error with derived field: operands could not be broadcast together with shapes
by Jean-Claude Passy 02 May '12

02 May '12

Hi all, I am trying to compare the energy between two datasets. I execute the following script that uses the derived field 'RelativeEnergy': -------------------------------------------------------------------------------------------------------------------------------------------------- def _RelativeEnergy(field, data): return (data["TotalEnergy"]-data.pf.parameters["TotalEnergy_ref"])/data.pf.parameters["TotalEnergy_ref"] add_field("RelativeEnergy", function=_RelativeEnergy) filen1 = 'helix/DD0000/CE0000' pf1 = load(filen1) region1 = pf1.h.region([0.5, 0.5, 0.5], [0.0, 0.0, 0.0], [1.0, 1.0, 1.0]) energy1 = region1["TotalEnergy"] filen2 = 'parallel/DD0000/CE0000' pf2 = load(filen2) region2 = pf2.h.region([0.5, 0.5, 0.5], [0.0, 0.0, 0.0], [1.0, 1.0, 1.0]) energy2 = region2["TotalEnergy"] # Adding reference energy into pf2 pf2.parameters["TotalEnergy_ref"] = energy1 delta_energy = region2["RelativeEnergy"] pc = PlotCollection(pf2) pc.add_slice("RelativeEnergy",2,center=[0.5, 0.5, 0.5]) pc.set_lim((0.4, 0.6, 0.4, 0.6)) pc.save('DeltaEnergy') -------------------------------------------------------------------------------------------------------------------------------------------------- I get the error pasted here: http://paste.yt-project.org/show/DPgMjV44JktpXExCljUj/ The variable delta_energy looks good (same type and size and energy1), but the problem arises when I try to take a slice of it. Any help much appreciated. Thanks a lot, JC

2 4

yt Google+ Hangout Tomorrow
by Sam Skillman 01 May '12

01 May '12

Hi all, Tomorrow we are going to try out a Google Hangout hosted by +yt. This is a chance for you to ask questions or iterate on a particular piece of analysis you'd like to perform. Also feel free to give feedback or just stop by and say hi to us. We'll be starting around 2pm EST tomorrow and there should be several of the yt developers hanging out. Depending on how it goes, we'll consider running one ever so often as a form of office hours or discussion of new capabilities. Anyways, if you have Google+ enabled and you head over and add +yt to your circles at https://plus.google.com/107728486871834552760/posts, you should see the hangout pop up in your feed. We'll also probably send out the link on the irc channel when we get started. I hope to see you there! Sam

3 3