- IMP-dev - salilab.org

Preparation for IMP 2.6.0 release
by Ben Webb 09 Mar '16

09 Mar '16

We're a little overdue for a new stable release. Since things are looking pretty good right now (most of the tests are passing) I plan to make a new stable release as soon as Charles is satisfied with the state of PMI. This will probably be within the next week, and certainly before the end of March. So now is a good time to fix minor annoyances, bugs, typos, etc. and not a great time to make large sweeping changes ;) Ben -- ben(a)salilab.org https://salilab.org/~ben/ "It is a capital mistake to theorize before one has data." - Sir Arthur Conan Doyle

1 0

IMP 2.5.0 release
by Ben Webb 01 Jul '15

01 Jul '15

IMP 2.5.0 is now available. This is a bugfix and minor update release. Most notably the documentation has been revised and updated; we now provide binary packages for Ubuntu 12.04 (Precise Pangolin) in addition to 14.04 (Trusty Tahr); a lot of obsolete and deprecated functionality has been removed; support for OpenMP and MPI is improved; and the IMP::base and IMP::kernel namespaces are now deprecated in favor of the top-level "IMP" namespace. See the http://integrativemodeling.org/2.5.0/doc/manual/changelog.html for the full change log. Check it out at http://integrativemodeling.org/ Ben -- ben(a)salilab.org http://salilab.org/~ben/ "It is a capital mistake to theorize before one has data." - Sir Arthur Conan Doyle

1 0

PMI commits mailing list
by Ben Webb 18 Jun '15

18 Jun '15

For those of you that follow PMI development, just FYI: there is a PMI-commits mailing list. Similar to the IMP-commits list, it gets an email every time code is pushed to the PMI repository at GitHub. (Full disclosure: it's actually been in existence for some time, but it never worked correctly - it should now.) To (un)subscribe, see https://salilab.org/mailman/listinfo/pmi-commits Ben -- ben(a)salilab.org http://salilab.org/~ben/ "It is a capital mistake to theorize before one has data." - Sir Arthur Conan Doyle

1 0

Preparation for 2.5.0 release
by Ben Webb 18 Jun '15

18 Jun '15

Since things look pretty stable right now, and there have been a bunch of changes since the last release, I'll branch for an IMP 2.5.0 release in the near future. Yell if you need me to wait on something. Ben -- ben(a)salilab.org http://salilab.org/~ben/ "It is a capital mistake to theorize before one has data." - Sir Arthur Conan Doyle

1 0

Re: [IMP-dev] [IMP-commits] [salilab/imp] 513c33: Add a support to write C-alpha residues and coarse...
by Seung Joong Kim 08 Jun '15

08 Jun '15

Can you upgrade "rmf_pdb" to get BEAD coordinates of the rmf file, as similar to what I did below? Of course, what I have done is not perfect - it is very preliminary without chain ID etc.. But at least generating some coordinates is much better than nothing saved. - SJ On 6/3/2015 4:26 PM, Notification of IMP commits wrote: > Branch: refs/heads/develop > Home: https://github.com/salilab/imp > Commit: 513c339597feef49bb0b532c21ca125f498e0d8c > https://github.com/salilab/imp/commit/513c339597feef49bb0b532c21ca125f498e0… > Author: procyon777(a)gmail.com <procyon777(a)gmail.com> > Date: 2015-06-03 (Wed, 03 Jun 2015) > > Changed paths: > M modules/atom/src/internal/pdb.cpp > > Log Message: > ----------- > Add a support to write C-alpha residues and coarsed-grained BEADs into a pdb file > > > > > _______________________________________________ > IMP-commits mailing list > https://salilab.org/mailman/listinfo/imp-commits -- Seung Joong Kim, Ph.D. Postdoctoral Scholar, Andrej Sali group Department of Bioengineering and Therapeutic Sciences University of California at San Francisco (UCSF) 1-217-649-2147 http://salilab.org/~sjkim

2 1

GPU cuda autocompilation in IMP
by Barak Raveh 30 Apr '15

30 Apr '15

I added an apparently operational cuda auto-build (any .cu file placed in src or src/internal should work). It seems to work fine but let me know if it causes havoc. Important note - the cuda build system hates the IMP namespace, specifically. Perhaps its some internal cuda macro. It means we cannot include <modulename>_config.h in any cuda files, for now, and need to add a bypass if we need to access an IMP method or object from cuda code. Barak

1 1

profiling imp code to boost your speed
by Barak Raveh 17 Apr '15

17 Apr '15

Hi all, Following Tuesday's discussion on slowness of IMP on some instances - all should be reminded of the powerful --cpu_profile IMP flag. Just add it to your command line, and IMP will generate a profiling file that you can view with google profiler as a tree: pprof --gv <your-binary-file> <the-profile-info-file-generated-by-cpu_profile-flag> it's almost certain you can at least double the performance of your app with a few simple steps. Likely more. Try to apply it to a long process (>60 seconds) so you will profile the interesting parts of your process, and not the overhead of loading files, etc. -- Barak

2 2

IMP 2.4.0 release
by Ben Webb 02 Mar '15

02 Mar '15

IMP 2.4.0 is now available. This is a bug fix and minor update release. Most notably Python 3 is now supported, and command line tools ("applications") are now part of regular IMP modules. See https://github.com/salilab/imp/releases/tag/2.4.0 for the full change log. Check it out at http://integrativemodeling.org/ Ben -- ben(a)salilab.org http://salilab.org/~ben/ "It is a capital mistake to theorize before one has data." - Sir Arthur Conan Doyle

1 0

IMP speed
by Davide Baù 10 Feb '15

10 Feb '15

Hi, we are experiencing some performance differences between different versions of IMP. IMP 2.0.1 seems to be much faster (up to 6 times faster) as compared to versions 2.1, 2.2 and 2.3.1 (compiled with the "Release" flag). Can you confirm this behavior or is it an issue affecting only our machines (both Mac and Linux)? Thanks, Davide

3 3

Python 3 support
by Ben Webb 12 Dec '14

12 Dec '14

Just FYI, latest IMP in git should now configure and build successfully if your 'python' binary is actually Python 3 (as with any nonstandard Python, you may need to set the cmake PYTHON_INCLUDE_DIR and PYTHON_LIBRARY variables to tell it where the associated headers and library are). Note that IMP itself doesn't fully work yet (there are still some issues with string typemaps, most likely because in Python 3 all strings are Unicode). The plan is to have our Python code run both in Python 2 and Python 3 (as opposed to maintaining a Python 2 version and a Python 3 version). This isn't too onerous since we don't support ancient versions of Python (older than 2.6) any more anyway. If you have Python code in IMP, I've modified it to work with both Python 2 and 3. You'll notice the major change is that "print" is no longer a statement in Python 3 (it's a function). So you have to say print("foo") rather than print "foo". I added "from __future__ import print_function" in IMP Python modules that use print so this will behave identically in Python 2 and 3. Other differences are fairly minor (e.g. 'xrange' is no more - use 'range' instead; some functions return iterators rather than lists in Python 3 so you have to wrap them with list() if you explicitly want the Python 2 behavior). Ben -- ben(a)salilab.org http://salilab.org/~ben/ "It is a capital mistake to theorize before one has data." - Sir Arthur Conan Doyle

1 0