Hi this question is mainly for Stephen,
I was working on using the ellipsoid branch of YT to do some analysis on the
3200 cube haloes found. I came across the same problem you described
earlier where calculating for tB_index I get nan sometimes, but this is
happening even after your fix. I dug through and found out that this
happens because pHOP found some halo that have only 1 particle.
This becomes problematic with my geometric ellipsoid, it requires at least 4
particles distributed in different planes. When the ellipsoid method is
trying to find a particle to satisfy the equation of an ellipse in 2D, it
will go through the list (of 1 element in this case), and find that it
returns nan for the length it calculates. So even though the function
nanargmax() is used to pick out biggest value of non-nan element index, with
a list of only 1 nan, it still gets selected and the method fails.
Currently my temporary solution is to tell the function that, if it
encounters a halo with less than 4 particles, to returns all 0s for the
ellipsoid attribute (A,B,C,e1x,y,z,tilt).
I was under the impression from our previous talk that there's a lowest
number threshold that pHOP requires to form a halo, I thought the number was
10 particles at least. Am I just remembering things wrong? It has been a
couple of years...
From
G.S.
Hi All,
yesterday I had a brain wave that it should be fairly simple to create
a new kind of data container that is a hybrid of existing ones, where
the mixing/cutting is done with boolean logic. I think I'm most of the
way there. Let me give you some examples of what this means.
>>> re1 = pf.h.region([0.5]*3, [0.4]*3, [0.6]*3)
>>> re2 = pf.h.region([0.5]*3, [0.5]*3, [0.6]*3)
>>> sp1 = pf.h.sphere([0.5]*3, 0.1)
>>> sp2 = pf.h.sphere([0.1]*3, 0.1)
This will give you the exact same region as re1:
>>> re = pf.h.boolean([re1, "OR'", re2])
This will give you the same as re2:
>>> re = pf.h.boolean([re1, "AND", re2])
This will give you the part of re1 that re2 does not cover, so re1
with a corner taken out:
>>> re = pf.h.boolean([re1, "NOT", re2])
Likewise, this will give you a box with a sphere cut out of it in the middle:
>>> re = pf.h.boolean([re1, "NOT", sp1])
This will combine the two disjoint spheres into one object:
>>> re = pf.h.boolean([sp1, "OR", sp2])
Nested logic should work using parentheses, where items in parentheses
are turned into new sub-hybrid regions before they are considered with
the others at a 'higher' level. This will give you a disjoint region,
with sp2 plus re2 with a corner cut out of it.
>>> re = pf.h.boolean([sp2, "OR", "(", re1, "NOT", re2, ")"])
This is all being done with cut_masks so it should work fairly
seamlessly. And so far, nearly everything seems to work. Except for
some reason that I don't fully grasp right now, I'm changing the
cut_masks of the starting regions, which is wrong. Here's an example
of what I mean. If I make a hybrid region:
In [1]: re1 = pf.h.region([0.5]*3, [0.4]*3, [0.6]*3)
In [2]: re2 = pf.h.region([0.5]*3, [0.5]*3, [0.6]*3)
In [3]: re = pf.h.boolean([re1, "AND", re2])
In [4]: re1['x'].shape
yt : [INFO ] 2011-10-29 10:00:25,002 Getting field x from 4
Out[4]: (2197,)
I'm seeing that re1 is being modified. Here's what I should see for
the shape of ['x']:
In [1]: re1 = pf.h.region([0.5]*3, [0.4]*3, [0.6]*3)
In [2]: re1['x'].shape
yt : [INFO ] 2011-10-29 10:14:04,730 Getting field x from 4
Out[2]: (17576,)
It's being changed to the dimensions of re2, which is what re is when
boolean is done.
I'm wondering if I could get a second pair of eyes on this, someone
who understands the subtleties of data_containers better than me. I've
pushed the current state of things to my branch.
https://bitbucket.org/sskory/yt/changeset/148e930ce903
Thanks to anyone who can spare the time, or for any comments on this
idea! Let me know if you have any questions.
--
Stephen Skory
s(a)skory.us
http://stephenskory.com/
510.621.3687 (google voice)
Hi, all--
I've notice that trailing comments are not longer ignored when parsing
parameter files. I have a parameter in many of my parameter files
that documents itself, as in
WriteBoundary = 0 //1=already written, 2=start writing
which recently started causing the parameter file to parse. Is this
an easy thing to fix in yt, or was this change done intentionally for
a broader reason, and I should write a sed script to pull this out? I
wouldn't bother asking if it weren't in an enormous number of files.
Thanks,
d.
--
Sent from my computer.
Hi all,
I think I have the column density thingy working. I've got slices,
projections and volume renders working (http://vimeo.com/31035296). I
have a few questions.
- Where in yt should I put this? Keeping the file in my work directory
I use it like this, below. yt/utilities? Better ideas?
from ColumnDensity import *
cdnumdens = ColumnDensity(pf, 'NumberDensity', [0.5, 0.5, 0.5])
def _CDNumberDensity(field, data, cd = cdnumdens):
return cd._build_derived_field(data)
add_field('CDNumberDensity', _CDNumberDensity)
- The function _build_derived_field, used directly above, is slightly
odd. What happens in it is I get the cell positions from data
(data['x'], etc..) and put those into the stuff that interpolates the
column density. I get a field based on these positions, and that's
what gets returned. However, and this shows what I don't understand,
for projections (as in not slices) one of the times this function gets
called as part of the add_field machinations data is a FieldDetector,
and calling data['x'] on them is not what I want. So I've added an
"if" that basically returns na.ones when data is a FieldDetector. What
would be a better option than that? The source as it stands is pasted
below.
http://paste.enzotools.org/show/1892/
Thanks for your thoughts!
--
Stephen Skory
s(a)skory.us
http://stephenskory.com/
510.621.3687 (google voice)
Hi, Everybody--
In order to get units to be converted properly for fields that do not
have units information in the parameter file, I've been subclassing
EnzoStaticOutput. However, for some reason this shorts out the load
function. For instance merely subclassing EnzoStaticOutput as follows
causes load to return None, instead of a parameter file. This changed
some time ago, but I've been ignoring it. Is there a way I can
continue subclassing the EnzoStaticOutput, so I have more control over
my units and not lose the load function?
Thanks,
d.
from yt.mods import *
class FileStaticOutputFourPi(EnzoStaticOutput):
def __init__(self,*args,**kwargs):
fpi = na.sqrt(4*na.pi)
EnzoStaticOutput.__init__(self,*args,conversion_override={'Density': 1,
'DivB': 1,
'x-velocity': 1,
'Total_Energy': 1},
parameter_override={'LengthUnits':
1,'LengthUnit': 1,
'Time': 1},
**kwargs)
>>> pf = load(path)
>>> print pf
None
--
Sent from my computer.
Hi all,
A few of us worked this past week on a couple yt projects and made what we
think is significant progress. Two of the items we focused on were testing
and parallelism.
For testing, we've broadened the test suite to include many more functions
and derived quantities. We now have 548 tests that include (off and
on-axis) slices, (off and on-axis) projections, phase distributions, halo
finding, volume rendering, and geometrical region cuts such as rectangular
solids, spheres, and disks. We use both plain and derived fields for these
tests so that it covers as many bases as possible. With this framework, we
are now able to keep a gold standard of the test results for any dataset,
then test later changes against this standard. These tests can test for
bitwise identicality or allow for some tolerance. For a full list of tests,
you can run python yt/tests/runall.py -l, and use --help to look at the
usage. We will soon be updating the documentation to provide more
information on how to set up the testing framework, but I think all of us
agree that this will make it much easier to test our changes to make sure
bugs have not crept in.
The second big change I'd like to talk about is the way we now handle
parallelism in yt. Previously, methods that employed parallelism through
MPI calls would first inherit from ParallelAnalysisInterface, which had
access to a ton of mpi functions that all work off of MPI.COMM_WORLD. In
our revamp we wanted to accomplish two things: 1) merge duplicate mpi calls
that were only different by the type of values they work on and do overall
cleanup. 2) Allow for nested levels of parallelism where two (or more)
separate communicators are able to use barriers and collective operations
such as allreduce. To do this, we worked in a two-step process. First we
took things like:
def _mpi_allsum(self, data):
def _mpi_Allsum_double(self, data):
def _mpi_Allsum_long(self, data):
def _mpi_allmax(self, data):
def _mpi_allmin(self, data):
and packed it into a single function:
def mpi_allreduce(self, data, dtype=None, op='sum'):
When a numpy array is passed to this new mpi_allreduce, dtype is determined
from the array properties. If the data is a dictionary, then it is passed
to mpi4py's allreduce function that acts on dictionaries. This greatly
reduced the number of lines in parallel_analysis_interface (1376 ==> 915),
even after adding in additional functionality.
The second step was bundling all of these functions into a new class called
Communicator. This Communicator object is initialized with an MPI
communicator that no longer is restricted to COMM_WORLD. Using this as the
fundamental MPI object, we then built a CommunicationSystem object that
manages these communicators. A global communication_system instance is
created, that is initialized with COMM_WORLD at the top of the system if the
environment is mpi4py-capable. If not, an empty communicator is created
that has passthroughs for all the mpi functions.
Using this new framework we are now able to take advantage of multiple
communicators. There are two use cases that we have implemented so far:
1) parallel_objects
parallel_objects is a method un parallel_analysis_interface.py for iterating
over a set of objects such that a group of processors work on each object.
This could be used, for example, to run N projections each with M
processors, allowing for a parallelism of NxM.
2) workgroups
workgoups allows users to set up multiple MPI communicators with a
non-uniform number of processors to each work on a separate task. This
capability lives within the ProcessorPool and Workgroup objects in
parallel_analysis_interface.py
These are just the first two that we tried out and we are very excited about
the new possibilities.
With these changes, there was one implementation change that has already
come up once in the mailing list. When you implement a new class that you'd
like to have access to the communication objects, you must first inherit
ParallelAnalysisInterface, and then make sure that __init__ makes a call to:
ParallelAnalysisInterface.__init__()
At that point, your new class will have access to the mpi calls through the
self.comm object. For example, to perform a reduction one would do:
self.comm.mpi_allreduce(my_data, op='sum')
As I said before, this will all be documented soon, but hopefully this will
help for now.
Sam, Britton, Cameron, Jeff, and Matt
Hi all,
I'm finding a problem with the new communicator stuff in
ParallelAnalysisInterface in r322fe24bb0cb. Specifically, .comm is
being initialized to None. With a short script like this:
~~~~~~~~~~~~
from yt.mods import *
from yt.utilities.parallel_tools.parallel_analysis_interface import \
ParallelAnalysisInterface
class humbug(ParallelAnalysisInterface):
def __init__(self):
pass
bah = humbug()
print bah.comm
~~~~~~~~~~~~
I get "None", which is not correct, I think. I've found that if I
modify things to look like this below in
parallel_analysis_interface.py, it works. But I don't think that's how
things are supposed to work. Could someone who understands what's
supposed to happen take a look at this? I suppose _grids and
_distributed would also suffer from the same problem. Thanks!
~~~~~~~~~~~~
class ParallelAnalysisInterface(object):
comm = communication_system.communicators[-1]
_grids = None
_distributed = None
def __init__(self):
self.comm = communication_system.communicators[-1]
self._grids = self.comm._grids
self._distributed = self.comm._distributed
~~~~~~~~~~~~~~
--
Stephen Skory
s(a)skory.us
http://stephenskory.com/
510.621.3687 (google voice)
Hi YT devs, I was asked by the sys admin at NICS about a part of YT that I'm
not very familiar with:
"Is the particle IO in YT that calls h5py spawned by multiple processors or
is it doing it serially?"
We're trying to solve a memory issue we're seeing in the parallelHF, and
Stephen has been making modifications that I'm currently trying out, but we
wanted to see if we can identify the source of the problem. I told the
admin that parallel projection works with each processor reading in only a
portion of the field variables, but I couldn't be sure about the particle
data, so I wanted to verify with people on the dev list.
From
G.S.
Enzo 2.1
Oct. 17, 2011
We are proud to announce the public release of Enzo version 2.1. New with this release is an extensive answer testing facility, additional physics capabilities, and AMR performance enhancements. New physics capabilities include isotropic and anisotropic heat conduction, X-ray ionization/heating, massive black hole particles, 4th-order accurate gravity solver, distributed stellar feedback, shock tracking, and improved molecular hydrogen chemistry. Performance improvements include Hilbert curve dynamic load balancing, faster ray tracing, and ray merging. We have improved Enzo’s inline analysis using the yt project, as well as user documentation for the entire Enzo codebase. We have also simplified the generation of multi-mesh cosmological initial conditions.
In addition to code upgrades, we have upgraded and reorganized our online repositories and documentation. The Enzo Project’s main website is now http://enzo-project.org/ . Links and information can be found here for both developers and general users. Our stable releases will continue to live on our Google code website: http://code.google.com/p/enzo/ . Users can clone a mercurial repository of our stable release code or download a tarball of the code (http://code.google.com/p/enzo/downloads/list ). A simple quick-start guide can also be found here: http://code.google.com/p/enzo/wiki/EnzoBootCamp . We will now support versioned documentation for all stable releases since Enzo 2.0 in addition to the development code on enzo-project.org. The documentation for Enzo 2.1 is located here: http://enzo-project.org/docs/2.1/.
Development will now move to our new Bitbucket repository where all users are encouraged to make contributions to the code by forking the main repository, located at https://bitbucket.org/enzo/enzo-dev . An explanation of new developer guidelines can be found here: http://enzo-project.org/docs/2.1/developer_guide/index.html . The current release of Enzo 2.1 features contributions, bug fixes and enhancements from an additional ten individuals (over the course of 2700 changesets) since Enzo 2.0, and through our new development infrastructure we are expanding our ability to solicit and review contributions.
Main website: http://enzo-project.org/
Mailing List: https://groups.google.com/forum/#!forum/enzo-users
Stable repository: http://code.google.com/p/enzo/
Tarball of Enzo 2.1.0: http://code.google.com/p/enzo/downloads/list
Enzo Boot Camp (a quick-start guide): http://code.google.com/p/enzo/wiki/EnzoBootCamp
Enzo 2.1.0 documentation: http://enzo-project.org/docs/2.1/
Developer’s repository: https://bitbucket.org/enzo/enzo-dev/overview
Developer’s guidlines: http://enzo-project.org/docs/2.1/developer_guide/index.html
The Enzo Development Team
Hi all,
I screwed up and pushed a bunch of changes to the main yt repository
that I didn't intend to. These include the following:
* Experimental changes to database handling. These were in the
branch 'yt'. Through a no-op merge I removed these from causing
problems. I hope that some day these changes will become part of the
main repository, but they definitely are not ready yet.
* Experimental changes in the named branch deliberate_fields
* Experimental changes in the named branch geometry_handling.
I apologize for this, but as long as you don't update to these named
branches, it shouldn't be too much of a problem. Hopefully they'll
all three get merged in eventually.
-Matt
