As noted below, there's going to be a development workshop for yt at
NCSA at University of Illinois Urbana-Champaign, on October 16 and 17.
This is going to be a dev workshop, but *everyone* is invited (whether
you use Blue Waters or not!). There isn't yet a set schedule, but I
hope we can spend time on solidifying support for newer frontends,
attaching to running simulations, and hopefully even converting the
n-body codes to forest-of-octree from the current implementation. I'm
also thinking it might be a good idea to block off an afternoon *just*
for converting any analysis modules or analysis code that works on yt
2 but not yt 3.
In the next few days I will have information about funding for
attendance; please write to me to let me know if you'd like travel
support. I can't guarantee anything yet, but I am hoping to have some
positive information soon.
Hope to see lots of people there!
---------- Forwarded message ----------
Date: Mon, Sep 1, 2014 at 3:16 PM
Subject: Blue Waters Advanced User Workshop and yt development workshop
To: Blue Waters Announcements <noreply(a)ncsa.illinois.edu>
We are pleased to announce a three-day, hands-on oriented advanced
user workshop about using Blue Waters to its fullest potential.
Representatives from Cray, Allinea, PGI and NVIDIA will be on-site to
present material and interact with attendees so bring your own code.
This work shop is targeting existing or recently allocated, advanced
users of Blue Waters.
In conjunction with the Blue Waters Advanced User workshop, we are
pleased to invite people who are interested in participating in
development of the yt project (yt-project.org) to a two-day hands-on
development workshop, focusing on parallel and in situ analysis,
stabilizing existing data format front-ends, and expanding the
capabilities of yt to take advantage of supercomputing resources.
Blue Waters Advanced User workshop
yt development workshop
We hope to see you here.
I apologize in advance for my naivety...
I asked this in IRC, and Kacper was kind enough to provide some insight,
but alas I was still having problems getting this to work. I know it
involves bookmarks and heads, but I'm still not as familiar with mercurial
as I am with git, so this all confuses my gitsense. Also, I know there are
others who have this issue - perhaps this should be on the docs page
Let's say I have my fork of yt wherein I already have a pending PR on BB.
I can't simply make changes to my fork and push them in the default
branch/bookmark because BB likes to automagically update PRs.
What is the recommended procedure for committing and pushing new changes to
my fork so that I can issue an additional PR? Do I create a bookmark, then
commit my changes, then push that bookmark to BB? If so, how then can I
issue a PR from that bookmark - I can't seem to find this on the BB
New issue 895: Can't load mass of stars from Gadget snapshots written to multiple files
I have this recurring problem when using snapshots written to multiple files with Gadget. I'm interested in following the total stellar mass at each time step. However when using multiple files i get the error: http://paste.yt-project.org/show/5107/ when trying to reference particular files. I redid my simulation writing to one file per snapshot and never had a problem referencing the mass of all the stellar particles. Attached you will find the snapshot written to one file (snapshot_025.hdf5) and one written to multiple files (in the snapdir_025).These are from the exact same simulation, the only difference being how many files the snapshots were distributed over. Both cases have been run a second time and the problem has still remained.
Members, if you are interested in taking on some leadership within yt,
please consider signing yourself to be a representative for one for the
five subcomponents: infrastructure, plotting, testing, documentation, and
analysis modules. If you have experience with one of these areas or are
interested in becoming more involved in planning and/or development,
consider yourself right for the position. An area should ideally be
represented by at least a two or three people. All that is required is
that someone from each subcomponent be present at a monthly hangout to
discuss the latest comings and goings. Additionally, if you are the first
person to sign up for one area, consider reaching out to a couple others
who you think would also be right for the position.
If you'd like to sign up, clone the YTEP repo (
https://bitbucket.org/yt_analysis/ytep) and add your name to YTEP-1776 and
either issue a PR or just push directly. When changes are pushed to the
repo, the document will get automatically rebuilt and will be viewable
New issue 894: Test suite does not detect load function breakage
This issue came up regarding PR 1190. Currently, if one breaks the code in such a way that the load function fails for certain datasets, the test suite will skip those datasets instead of reporting a failure.
IME, breaking the _is_valid methods is really easy to do when doing frontend work, and it be nice if the answer tests reported a failure when this happens.
As Nathan suggested in IRC, the requires_ds decorator probably just needs to be made smarter.
New issue 893: Feature request: specify xyz or zyx order in load_octree
The load_octree currently makes assumptions about how to map the 1-d array of quantities into three dimensions. I think it varies fastest along z, then along y, then along x. It would be nice to be able to specify this order using an argument that could be set to either e.g. ``xyz`` or ``zyx``.
I'd like to nominate Chris Malone for member status. Chris is the one who
got me introduced to yt (full disclosure, I was his thesis advisor). Chris
has actively hacked on the boxlib stuff (including a PR awaiting two more
approvals right now), as well as a number of fixes throughout that you can
see in the hg logs, and has taken an active role in answering questions on
the yt-user's list.
Dept. of Physics & Astronomy • Stony Brook University • Stony Brook, NY
I'd like to nominate Kenza Arraki for membership.
Kenza has acted as the representative for the NMSU-ART code, and has
spent a considerably amount of time and effort working on making sure
it's correctly set up, and usable for the folks using it. I think she
would be an excellent liaison for NMSU-ART, and a great addition to
the membership ranks.