<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Thank you, Kacper - that was fast (!), and it works nicely!<br>
<br>
For general information -- here's the procedure I used to apply your
PR to my local copy. Is this the straightest path to doing that?<br>
<br>
Look at your PR page,<br>
go to Commits tab,<br>
hover over the "
<meta http-equiv="content-type" content="text/html;
charset=windows-1252">
<a class="hash execute"
href="https://bitbucket.org/xarthisius/yt/commits/207f56eb6859d529ac4f6170720317c69ebdfad7?at=yt">207f56e</a>"
commit link and Copy Link Location =><br>
<a
href="https://bitbucket.org/xarthisius/yt/commits/207f56eb6859d529ac4f6170720317c69ebdfad7?at=yt"><a class="moz-txt-link-freetext" href="https://bitbucket.org/xarthisius/yt/commits/207f56eb6859d529ac4f6170720317c69ebdfad7?at=yt">https://bitbucket.org/xarthisius/yt/commits/207f56eb6859d529ac4f6170720317c69ebdfad7?at=yt</a></a><br>
edit that URL to identify the source repo - yours:<br>
<a class="moz-txt-link-freetext" href="https://bitbucket.org/xarthisius/yt">https://bitbucket.org/xarthisius/yt</a><br>
<br>
Then from the command line, in my yt source tree, compare your
repo with what I've got:<br>
hg incoming -G <a class="moz-txt-link-freetext" href="https://bitbucket.org/xarthisius/yt">https://bitbucket.org/xarthisius/yt</a><br>
or better, to see just the changes that are related to your new
PR commit,<br>
hg incoming -G -r
<meta http-equiv="content-type" content="text/html;
charset=windows-1252">
<span class="hash execute">207f56e </span><a class="moz-txt-link-freetext" href="https://bitbucket.org/xarthisius/yt">https://bitbucket.org/xarthisius/yt</a><br>
and indeed they look reasonable, so:<br>
hg pull -r <span class="hash execute">207f56e </span><a class="moz-txt-link-freetext" href="https://bitbucket.org/xarthisius/yt">https://bitbucket.org/xarthisius/yt</a><br>
hg up<br>
<br>
and then from the .../src/yt-hg top of the source tree<br>
python setup.py develop<br>
<br>
Right? It's simple enough to do it this way - just posting this in
case there's some even simpler trick that I'm missing.<br>
<br>
<br>
<div class="moz-cite-prefix">On 3/22/16 11:05 AM, Kacper Kowalik
wrote:<br>
</div>
<blockquote cite="mid:56F16D55.7050703@gmail.com" type="cite">On
03/21/2016 06:07 PM, Stuart Levy wrote:
<br>
<blockquote type="cite">Hello yt people,
<br>
<br>
I'm trying to use MPI on a simple Enzo volume-rendering script,
using
<br>
the latest yt33 dev from bitbucket, and am getting a puzzling
<br>
behavior. (MPI so as to reduce the load time.)
<br>
<br>
Symptom: the datatype of the Scene.last_render field is
inconsistent:
<br>
it's YTImageArray when successful, but just plain YTArray when
using MPI
<br>
on the right (wrong) data, and so sc.save() fails. It's easy
to work
<br>
around and save the image another way, but makes me worried that
<br>
something else is wrong, or that I'm misusing yt.
<br>
<br>
It is somehow data-dependent. The error doesn't happen when I
use the
<br>
tiny enzo dataset in .../yt-hg/tests/DD0010/moving7_0010, nor
the
<br>
enzo_tiny_cosmology sample (DD0043). But it *does* happen when
I use a
<br>
somewhat-larger-but-not-huge simulation from John Wise - the
<br>
low-resolution PopIII-star simulation he'd given us last year.
A
<br>
single timestep from it is ~46MB and has only a few dozen grids.
<br>
<br>
I could post the data somewhere if John Wise would be OK with
that - John?
<br>
<br>
Demo code in:
<br>
<a class="moz-txt-link-freetext" href="http://paste.yt-project.org/show/6343/">http://paste.yt-project.org/show/6343/</a>
<br>
And its output when run under mpirun:
<br>
<a class="moz-txt-link-freetext" href="http://paste.yt-project.org/show/6344/">http://paste.yt-project.org/show/6344/</a>
<br>
<br>
The one change to the yt code from 'tip' (bed01b2c838c) is this
diagnostic:
<br>
diff -r bed01b2c838c yt/visualization/volume_rendering/scene.py
<br>
--- a/yt/visualization/volume_rendering/scene.py Mon Mar 21
14:15:54
<br>
2016 -0700
<br>
+++ b/yt/visualization/volume_rendering/scene.py Mon Mar 21
18:01:20
<br>
2016 -0500
<br>
@@ -284,6 +284,7 @@
<br>
self.render()
<br>
<br>
mylog.info("Saving render %s", fname)
<br>
+ print "about to write_png() - Scene.last_render",
<br>
type(self.last_render), "shape",self.last_render.shape, " =>
",
<br>
dir(self.last_render)
<br>
self.last_render.write_png(fname,
sigma_clip=sigma_clip)
<br>
</blockquote>
<br>
Hi Stuart,
<br>
I've issued PR 2066 [1] that should fix this.
<br>
Cheers,
<br>
Kacper
<br>
<br>
[1] <a class="moz-txt-link-freetext" href="https://bitbucket.org/yt_analysis/yt/pull-requests/2066/">https://bitbucket.org/yt_analysis/yt/pull-requests/2066/</a>
<br>
_______________________________________________
<br>
yt-users mailing list
<br>
<a class="moz-txt-link-abbreviated" href="mailto:yt-users@lists.spacepope.org">yt-users@lists.spacepope.org</a>
<br>
<a class="moz-txt-link-freetext" href="http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org">http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org</a>
<br>
</blockquote>
<br>
</body>
</html>