Merged revisions 14372,14374,14376,14386-14388,14390-14391 via svnmerge from
svn+ssh://bugman@.../svn/relax/1.3
................
r14372 | bugman | 2011-08-12 12:40:58 +0200 (Fri, 12 Aug 2011) | 14 lines
Merged the changes to the CHANGES file from the 1.3.11 tag.
The command used was:
svn merge -r14370:14371 svn+ssh://bugman@.../svn/relax/tags/1.3.11
.....
r14371 | bugman | 2011-08-12 12:36:47 +0200 (Fri, 12 Aug 2011) | 3 lines
Changed paths:
M /tags/1.3.11/docs/CHANGES
Updated the CHANGES file for relax version 1.3.11.
.....
................
r14374 | bugman | 2011-08-12 14:26:44 +0200 (Fri, 12 Aug 2011) | 3 lines
Fix to the structure.read_xyz user function docstring so that the user manual can be compiled.
................
r14376 | bugman | 2011-08-12 14:47:07 +0200 (Fri, 12 Aug 2011) | 3 lines
Updated the release checklist to include the bundling of bmrblib (http://gna.org/projects/bmrblib).
................
r14386 | bugman | 2011-08-16 11:40:55 +0200 (Tue, 16 Aug 2011) | 3 lines
Initialized merge tracking via "svnmerge" with revisions "1-5969" from
svn+ssh://bugman@.../svn/relax/branches/multi_processor_merge
................
r14387 | bugman | 2011-08-16 13:41:16 +0200 (Tue, 16 Aug 2011) | 1608 lines
Merged revisions 7709-7727,7729,7732,7734-7738,7740,7745-7783,7794,7808-7836,7838-7851,7853-7865,7875,9408,9448-9453,9997,10119-10156,10159-10177,10189-10196,10198-10203,10758,10760-10765,10790-10793,10830-10837,10840,10844,10861,10864,10867-10870,12420-12421 via svnmerge from
svn+ssh://bugman@.../svn/relax/branches/multi_processor_merge
........
r7709 | bugman | 2008-10-15 21:40:06 +0200 (Wed, 15 Oct 2008) | 18 lines
Manually merged r3208 from the multi_processor branch.
The command used was:
svn merge -r3207:3208 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
This marks the start of an interesting porting experience!
.....
r3208 | varioustoxins | 2007-03-16 12:47:30 +0100 (Fri, 16 Mar 2007) | 3 lines
Changed paths:
M /branches/multi_processor/relax
initial work to add multiprocessor startup and refactoring of main Relax
class
.....
........
r7710 | bugman | 2008-10-15 21:46:10 +0200 (Wed, 15 Oct 2008) | 18 lines
Manually ported r3210 from the multi_processor branch.
The command used was:
svn merge -r3209:3210 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3210 | varioustoxins | 2007-03-16 15:22:49 +0100 (Fri, 16 Mar 2007) | 2 lines
Changed paths:
A /branches/multi_processor/multi
A /branches/multi_processor/multi/__init__.py
A /branches/multi_processor/multi/mpi4py_processor.py
A /branches/multi_processor/multi/uni_processor.py
M /branches/multi_processor/relax
updates to get first mpi initialization working
.....
........
r7711 | bugman | 2008-10-15 21:48:43 +0200 (Wed, 15 Oct 2008) | 3 lines
Added a change which occurred in r3208 of the multi_processor branch but was somehow not ported.
........
r7712 | bugman | 2008-10-15 21:51:39 +0200 (Wed, 15 Oct 2008) | 15 lines
Manually merge r3235 from the multi_processor branch.
The command used was:
svn merge -r3234:3235 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3235 | varioustoxins | 2007-03-19 11:57:19 +0100 (Mon, 19 Mar 2007) | 4 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
consoles and scripts now exit cleanly (override sys.exit) interactive
console works!
.....
........
r7713 | bugman | 2008-10-15 21:54:17 +0200 (Wed, 15 Oct 2008) | 16 lines
Manually merged r3236 from the multi_processor branch.
The command used was:
svn merge -r3235:3236 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
This contains a bit more than the log message indicates, hence why this was ported!
.....
r3236 | varioustoxins | 2007-03-19 11:58:24 +0100 (Mon, 19 Mar 2007) | 3 lines
Changed paths:
M /branches/multi_processor/relax
marked more threading code for reletion
.....
........
r7714 | bugman | 2008-10-15 21:55:41 +0200 (Wed, 15 Oct 2008) | 15 lines
Manually ported r3237 from the multi_processor branch.
The command used was:
svn merge -r3236:3237 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3237 | varioustoxins | 2007-03-19 16:13:15 +0100 (Mon, 19 Mar 2007) | 2 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/relax
implimentation of basic command structure and simplistic processor name command
.....
........
r7715 | bugman | 2008-10-15 22:04:11 +0200 (Wed, 15 Oct 2008) | 16 lines
Manually ported r3238 from the multi_processor branch.
The command used was:
svn merge -r3237:3238 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3238 | varioustoxins | 2007-03-19 23:23:33 +0100 (Mon, 19 Mar 2007) | 3 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/relax
updates to create a better inter process command passing mechanism
and initiate relax in what might possibly be a sane state!
.....
........
r7716 | bugman | 2008-10-15 22:05:14 +0200 (Wed, 15 Oct 2008) | 16 lines
Manually ported r3239 from the multi_processor branch.
The command used was:
svn merge -r3238:3239 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3239 | varioustoxins | 2007-03-22 09:20:55 +0100 (Thu, 22 Mar 2007) | 3 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/relax
multi processor fixes for proper command queing with optional
load balancing iexception support and segmented results
.....
........
r7717 | bugman | 2008-10-15 22:10:31 +0200 (Wed, 15 Oct 2008) | 17 lines
Partially ported r3241 from the multi_processor branch.
The command used was:
svn merge -r3240:3241 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
The changes to maths_fns/mf.py didn't need to be ported across.
.....
r3241 | varioustoxins | 2007-03-27 15:35:16 +0200 (Tue, 27 Mar 2007) | 2 lines
Changed paths:
M /branches/multi_processor/maths_fns/mf.py
M /branches/multi_processor/multi/mpi4py_processor.py
installation of simplistic remote minimisation command
.....
........
r7718 | bugman | 2008-10-15 22:40:20 +0200 (Wed, 15 Oct 2008) | 16 lines
Manually ported r3242 from the multi_processor branch.
The command used was:
svn merge -r3241:3242 svn+ssh://bugman@.../svn/relax/branches/multi_processor/specific_fns/model_free.py specific_fns/model_free/mf_minimise.py
This was a bugger!
....
r3242 | varioustoxins | 2007-03-27 17:04:02 +0200 (Tue, 27 Mar 2007) | 3 lines
Changed paths:
M /branches/multi_processor/specific_fns/model_free.py
changes to add minimise command processing for multiprocessing
....
........
r7719 | bugman | 2008-10-15 22:53:04 +0200 (Wed, 15 Oct 2008) | 23 lines
Manually ported r3243 from the multi_processor branch.
The 2 commands used were:
svn merge -r3242:3243 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
svn merge -r3242:3243 svn+ssh://bugman@.../svn/relax/branches/multi_processor/specific_fns/model_free.py specific_fns/model_free/mf_minimise.py
This code is becoming horribly broken. This can't be fixed yet as it will make porting impossible!
.....
r3243 | varioustoxins | 2007-03-29 11:45:22 +0200 (Thu, 29 Mar 2007) | 4 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/uni_processor.py
M /branches/multi_processor/prompt/interpreter.py
M /branches/multi_processor/relax
M /branches/multi_processor/specific_fns/model_free.py
First fully working multi branch with both uniprocessor and mpi4py support
communication overhead for 18 residues (test_short.py from chris) with
in memory io ~25%
.....
........
r7720 | bugman | 2008-10-15 22:55:11 +0200 (Wed, 15 Oct 2008) | 17 lines
Manually ported r3246 from the multi_processor branch.
The command used was:
svn merge -r3245:3246 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3246 | varioustoxins | 2007-03-29 19:43:49 +0200 (Thu, 29 Mar 2007) | 3 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
A /branches/multi_processor/multi/processor.py
M /branches/multi_processor/multi/uni_processor.py
corrections to remove unexpected calls of from mpi4py import MPI in
mpi4py_processor from uniprocessor leading to an unexpected requirement of mpi4py
.....
........
r7721 | bugman | 2008-10-15 22:57:55 +0200 (Wed, 15 Oct 2008) | 19 lines
Manually ported r3248 from the multi_processor branch.
The 2 commands used were:
svn merge -r3247:3248 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
svn merge -r3247:3248 svn+ssh://bugman@.../svn/relax/branches/multi_processor/specific_fns/model_free.py specific_fns/model_free/mf_minimise.py
.....
r3248 | varioustoxins | 2007-03-29 21:47:17 +0200 (Thu, 29 Mar 2007) | 4 lines
Changed paths:
A /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/specific_fns/model_free.py
more modifications to remove accidental dependances on mpi4py in uni_processor
creation of command module
.....
........
r7722 | bugman | 2008-10-15 23:04:33 +0200 (Wed, 15 Oct 2008) | 19 lines
Manually ported r3249 from the multi_processor branch.
The 2 commands used were:
svn merge -r3248:3249 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
svn merge -r3248:3249 svn+ssh://bugman@.../svn/relax/branches/multi_processor/specific_fns/model_free.py specific_fns/model_free/mf_minimise.py
.....
r3249 | varioustoxins | 2007-03-30 09:00:39 +0200 (Fri, 30 Mar 2007) | 4 lines
Changed paths:
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/uni_processor.py
M /branches/multi_processor/specific_fns/model_free.py
more work to remove specious calls to module mpi4py when starting
uniprocessor version feedback prints correctly
.....
........
r7723 | bugman | 2008-10-15 23:05:57 +0200 (Wed, 15 Oct 2008) | 16 lines
Manually ported r3250 from the multi_processor branch.
The command used was:
svn merge -r3249:3250 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3250 | varioustoxins | 2007-03-30 09:32:44 +0200 (Fri, 30 Mar 2007) | 3 lines
Changed paths:
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
Minor name changes on MF memo slave and result commands
.....
........
r7724 | bugman | 2008-10-15 23:09:28 +0200 (Wed, 15 Oct 2008) | 17 lines
Manually ported r3251 from the multi_processor branch.
The 2 commands used were:
svn merge -r3250:3251 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
svn merge -r3250:3251 svn+ssh://bugman@.../svn/relax/branches/multi_processor/specific_fns/model_free.py specific_fns/model_free/mf_minimise.py
.....
r3251 | varioustoxins | 2007-03-30 09:53:32 +0200 (Fri, 30 Mar 2007) | 2 lines
Changed paths:
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/specific_fns/model_free.py
MF class name fixes also now works on python 2.3.3
.....
........
r7725 | bugman | 2008-10-15 23:10:22 +0200 (Wed, 15 Oct 2008) | 17 lines
Manually ported r3252 from the multi_processor branch.
The command used was:
svn merge -r3251:3252 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3252 | varioustoxins | 2007-03-30 09:57:54 +0200 (Fri, 30 Mar 2007) | 2 lines
Changed paths:
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/multi/uni_processor.py
applied GPL v2 licenses
.....
........
r7726 | bugman | 2008-10-15 23:12:00 +0200 (Wed, 15 Oct 2008) | 17 lines
Manually ported r3253 from the multi_processor branch.
The command used was:
svn merge -r3252:3253 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3253 | varioustoxins | 2007-03-30 10:58:03 +0200 (Fri, 30 Mar 2007) | 2 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/multi/uni_processor.py
M /branches/multi_processor/relax
'classification' of Processor classes
.....
........
r7727 | bugman | 2008-10-15 23:12:52 +0200 (Wed, 15 Oct 2008) | 15 lines
Manually ported r3254 from the multi_processor branch.
The command used was:
svn merge -r3253:3254 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3254 | varioustoxins | 2007-03-30 11:26:11 +0200 (Fri, 30 Mar 2007) | 2 lines
Changed paths:
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/uni_processor.py
corrections to keyword passing to dict.update()
.....
........
r7729 | bugman | 2008-10-15 23:32:04 +0200 (Wed, 15 Oct 2008) | 23 lines
Manually ported r3261 from the multi_processor branch.
The 2 commands used were:
svn merge -r3260:3261 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
svn merge -r3260:3261 svn+ssh://bugman@.../svn/relax/branches/multi_processor/specific_fns/model_free.py specific_fns/model_free/mf_minimise.py
.....
r3261 | varioustoxins | 2007-04-07 22:11:21 +0200 (Sat, 07 Apr 2007) | 4 lines
Changed paths:
M /branches/multi_processor/generic_fns/minimise.py
A /branches/multi_processor/multi/PrependStringIO.py
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/prompt/minimisation.py
M /branches/multi_processor/specific_fns/model_free.py
Major update - monte carlo simulations supported (trivially)
also implimented: chunky queues, exception throwing on slaves, and rank
annotated output. note uniprocessor support currently broken.
.....
........
r7732 | bugman | 2008-10-16 00:15:01 +0200 (Thu, 16 Oct 2008) | 19 lines
Merged all that could be merged from r3266.
The 2 commands used were:
svn merge -r3265:3266 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
svn merge -r3265:3266 svn+ssh://bugman@.../svn/relax/branches/multi_processor/specific_fns/model_free.py specific_fns/model_free/mf_minimise.py
The grid search code was skipped as it is in minfx (https://gna.org/projects/minfx/).
.....
r3266 | varioustoxins | 2007-04-12 18:48:28 +0200 (Thu, 12 Apr 2007) | 2 lines
Changed paths:
M /branches/multi_processor/minimise/grid.py
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/specific_fns/model_free.py
partial integration of distributed grid code (everything broken)
.....
........
r7734 | bugman | 2008-10-16 00:17:30 +0200 (Thu, 16 Oct 2008) | 20 lines
Merged all that could be merged from r3267.
The 2 commands used were:
svn merge -r3266:3267 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
svn merge -r3266:3267 svn+ssh://bugman@.../svn/relax/branches/multi_processor/specific_fns/model_free.py specific_fns/model_free/mf_minimise.py
The grid search code was skipped as it is in minfx (https://gna.org/projects/minfx/).
.....
r3267 | varioustoxins | 2007-04-17 00:54:48 +0200 (Tue, 17 Apr 2007) | 2 lines
Changed paths:
M /branches/multi_processor/minimise/generic.py
M /branches/multi_processor/minimise/grid.py
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/specific_fns/model_free.py
fully operative grid search (uniprocessor broken)
.....
........
r7735 | bugman | 2008-10-16 00:21:57 +0200 (Thu, 16 Oct 2008) | 19 lines
Manually ported r3268 from the multi_processor branch.
The 2 commands used were:
svn merge -r3267:3268 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
svn merge -r3267:3268 svn+ssh://bugman@.../svn/relax/branches/multi_processor/specific_fns/model_free.py specific_fns/model_free/mf_minimise.py
.....
r3268 | varioustoxins | 2007-04-17 14:39:25 +0200 (Tue, 17 Apr 2007) | 2 lines
Changed paths:
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/multi/uni_processor.py
M /branches/multi_processor/specific_fns/model_free.py
almost feature complete mpi4py and uni processor versions of relax
.....
........
r7736 | bugman | 2008-10-16 00:24:12 +0200 (Thu, 16 Oct 2008) | 32 lines
Manually ported r3269, r3270, and r3271 from the multi_processor branch.
The command used was:
svn merge -r3268:3271 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3271 | varioustoxins | 2007-04-23 13:06:57 +0200 (Mon, 23 Apr 2007) | 3 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
threaded results output to _TRY_ to avoid starving multiprocessor while
waiting for stdio
------------------------------------------------------------------------
r3270 | varioustoxins | 2007-04-21 03:18:39 +0200 (Sat, 21 Apr 2007) | 2 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
forgot to remove sys.exit used in testing
------------------------------------------------------------------------
r3269 | varioustoxins | 2007-04-21 03:17:26 +0200 (Sat, 21 Apr 2007) | 2 lines
Changed paths:
M /branches/multi_processor/multi/PrependStringIO.py
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
batched retrun results and better sys.exit and exception handling behaviour
.....
........
r7737 | bugman | 2008-10-16 00:26:21 +0200 (Thu, 16 Oct 2008) | 19 lines
Manually ported r3274 from the multi_processor branch.
The command used was:
svn merge -r3273:3274 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3274 | varioustoxins | 2007-04-27 23:58:18 +0200 (Fri, 27 Apr 2007) | 2 lines
Changed paths:
M /branches/multi_processor/minimise/generic.py
M /branches/multi_processor/multi/PrependStringIO.py
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/relax
correction to master slave output prefixes
.....
........
r7738 | bugman | 2008-10-16 00:33:21 +0200 (Thu, 16 Oct 2008) | 106 lines
Manually merged the last of the multi_processor branch.
The command used was:
svn merge -r3274:3296 svn+ssh://bugman@.../svn/relax/branches/multi_processor .
.....
r3296 | varioustoxins | 2007-05-16 23:46:21 +0200 (Wed, 16 May 2007) | 2 lines
Changed paths:
M /branches/multi_processor/multi/processor.py
more documentation strings (completed if buggy for multi.processor)
------------------------------------------------------------------------
r3295 | varioustoxins | 2007-05-15 23:12:10 +0200 (Tue, 15 May 2007) | 2 lines
Changed paths:
M /branches/multi_processor/multi/multi_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/specific_fns/model_free.py
minor improvements and much documentation of the Processor class
------------------------------------------------------------------------
r3283 | varioustoxins | 2007-05-08 21:45:48 +0200 (Tue, 08 May 2007) | 3 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/multi_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/relax
final separation between the multi framework and relax and initial
documentation of multi.processor
------------------------------------------------------------------------
r3280 | varioustoxins | 2007-05-04 01:07:07 +0200 (Fri, 04 May 2007) | 3 lines
Changed paths:
D /branches/multi_processor/multi/PrependStringIO.py
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/multi_processor.py
A /branches/multi_processor/multi/prependStringIO.py (from /branches/multi_processor/multi/PrependStringIO.py:3275)
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/multi/uni_processor.py
threaded multi processor support written, multi is now beta!
multi now supports three substrates: mpi4py, uniprocessor, and threads
------------------------------------------------------------------------
r3279 | varioustoxins | 2007-05-02 00:23:25 +0200 (Wed, 02 May 2007) | 3 lines
Changed paths:
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
A /branches/multi_processor/multi/multi_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/multi/uni_processor.py
major restructuring to create generic multiprocessor framework legacy
string based exception handling supported
------------------------------------------------------------------------
r3278 | varioustoxins | 2007-05-01 00:24:05 +0200 (Tue, 01 May 2007) | 3 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/multi/uni_processor.py
M /branches/multi_processor/relax
correct use of -n command line option, work to give general processor
base class and simple processor implimentations
------------------------------------------------------------------------
r3277 | varioustoxins | 2007-04-30 15:40:54 +0200 (Mon, 30 Apr 2007) | 3 lines
Changed paths:
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/multi/uni_processor.py
M /branches/multi_processor/relax
Addition of processor based intro string, correction of error in
Application_callback exceptio_handler named error_handler
------------------------------------------------------------------------
r3276 | varioustoxins | 2007-04-29 23:20:42 +0200 (Sun, 29 Apr 2007) | 4 lines
Changed paths:
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
M /branches/multi_processor/multi/uni_processor.py
M /branches/multi_processor/relax
modifications to make the multi hierachy indepndant of relax; multi can
now be used to run other multiprocessor tasks (uni processor and mpi
work)
------------------------------------------------------------------------
r3275 | varioustoxins | 2007-04-28 19:59:04 +0200 (Sat, 28 Apr 2007) | 2 lines
Changed paths:
M /branches/multi_processor/minimise/generic.py
M /branches/multi_processor/multi/PrependStringIO.py
M /branches/multi_processor/multi/commands.py
M /branches/multi_processor/multi/mpi4py_processor.py
M /branches/multi_processor/multi/processor.py
improved output and exception handling
.....
........
r7740 | bugman | 2008-10-16 00:42:05 +0200 (Thu, 16 Oct 2008) | 3 lines
Import fixes and clean ups and addition of module docstrings.
........
r7745 | bugman | 2008-10-16 13:30:33 +0200 (Thu, 16 Oct 2008) | 5 lines
Bug fix for the Uni_processor initialisation.
The std_stdio_capture() method has no rank arg.
........
r7746 | bugman | 2008-10-16 13:46:00 +0200 (Thu, 16 Oct 2008) | 5 lines
Clean up of the module docstring.
This includes whitespace fixes, punctuation fixes, spelling fixes, and fixes for the epydoc format.
........
r7747 | bugman | 2008-10-16 16:01:15 +0200 (Thu, 16 Oct 2008) | 6 lines
Big reformatting of all the docstrings.
This includes much tidying up and making the formatting of the text consistent, many spelling fixes,
epydoc fixes, etc.
........
r7748 | bugman | 2008-10-16 16:04:35 +0200 (Thu, 16 Oct 2008) | 3 lines
Shifted all functions to the top of the module and ordered them alphabetically.
........
r7749 | bugman | 2008-10-16 16:07:46 +0200 (Thu, 16 Oct 2008) | 3 lines
Ordered all the classes alphabetically.
........
r7750 | bugman | 2008-10-16 16:29:21 +0200 (Thu, 16 Oct 2008) | 5 lines
Massive set of white space fixes!!!
This was a huge clean up!
........
r7751 | bugman | 2008-10-16 16:31:16 +0200 (Thu, 16 Oct 2008) | 3 lines
Shifted the class Null_result_command to after Result_command due to it being a subclass.
........
r7752 | bugman | 2008-10-16 16:38:23 +0200 (Thu, 16 Oct 2008) | 5 lines
Sorted all class methods alphabetically.
This will really help with code navigation and debugging.
........
r7753 | bugman | 2008-10-16 16:39:36 +0200 (Thu, 16 Oct 2008) | 3 lines
Spelling fix for a function name.
........
r7754 | bugman | 2008-10-16 16:48:02 +0200 (Thu, 16 Oct 2008) | 5 lines
Big clean up of the relax module.
This was quite messy, and should be good enough for merging once all the FIXMEs are eliminated.
........
r7755 | bugman | 2008-10-16 16:51:07 +0200 (Thu, 16 Oct 2008) | 3 lines
Bug fix for the get_intro_string() method, intro_string is no longer a class variable.
........
r7756 | bugman | 2008-10-16 16:55:08 +0200 (Thu, 16 Oct 2008) | 3 lines
Added some clarifying comments.
........
r7757 | bugman | 2008-10-16 16:55:50 +0200 (Thu, 16 Oct 2008) | 3 lines
Added a docstring to print_stats().
........
r7758 | bugman | 2008-10-16 16:56:54 +0200 (Thu, 16 Oct 2008) | 5 lines
Converted the profile_flag variable to a bool.
Profiling is currently broken!
........
r7759 | bugman | 2008-10-16 16:58:04 +0200 (Thu, 16 Oct 2008) | 3 lines
Comment fix.
........
r7760 | bugman | 2008-10-16 16:59:25 +0200 (Thu, 16 Oct 2008) | 3 lines
Changed the processor type print out in the intro string.
........
r7761 | bugman | 2008-10-16 17:05:39 +0200 (Thu, 16 Oct 2008) | 3 lines
Placed the 'mode' variable into the class instance namespace.
........
r7762 | bugman | 2008-10-16 17:12:39 +0200 (Thu, 16 Oct 2008) | 5 lines
Bug fix and clean up of the multi-processor type and number command line args.
A number of command line args were failing because of this bug.
........
r7763 | bugman | 2008-10-16 17:13:23 +0200 (Thu, 16 Oct 2008) | 3 lines
Removed some unused code.
........
r7764 | bugman | 2008-10-16 17:19:12 +0200 (Thu, 16 Oct 2008) | 3 lines
sys.argv does not need to be passed into arguments(). The optparse module handles all of this.
........
r7765 | bugman | 2008-10-16 17:20:25 +0200 (Thu, 16 Oct 2008) | 3 lines
Removed some unused code (which appears to be old debugging code).
........
r7766 | bugman | 2008-10-16 17:31:43 +0200 (Thu, 16 Oct 2008) | 3 lines
Spacing fixes in the multi-processor import.
........
r7767 | bugman | 2008-10-16 17:34:41 +0200 (Thu, 16 Oct 2008) | 3 lines
Clean up of the disassemble_result() method.
........
r7768 | bugman | 2008-10-16 17:37:30 +0200 (Thu, 16 Oct 2008) | 3 lines
Formatting clean ups and a bug fix for the 'self.param_types' variable now named 'model_type'.
........
r7769 | bugman | 2008-10-16 17:39:20 +0200 (Thu, 16 Oct 2008) | 3 lines
Fixed the indentation in the disassemble_result() method.
........
r7770 | bugman | 2008-10-16 17:41:18 +0200 (Thu, 16 Oct 2008) | 3 lines
Shifted the disassemble_result() method to its alphabetically sorted position.
........
r7771 | bugman | 2008-10-16 17:46:12 +0200 (Thu, 16 Oct 2008) | 3 lines
Added contents to the multi package initialisation.
........
r7772 | bugman | 2008-10-16 17:56:48 +0200 (Thu, 16 Oct 2008) | 5 lines
Shifted the model-free multi-processor command classes and cleaned up the rest of the command module.
The model-free code has been shifted into the model-free 'specific_fns' package.
........
r7773 | bugman | 2008-10-16 17:57:39 +0200 (Thu, 16 Oct 2008) | 3 lines
Module rename.
........
r7774 | bugman | 2008-10-16 17:58:11 +0200 (Thu, 16 Oct 2008) | 3 lines
Updated the package __all__ list.
........
r7775 | bugman | 2008-10-16 18:00:06 +0200 (Thu, 16 Oct 2008) | 3 lines
Import fix.
........
r7776 | bugman | 2008-10-16 18:03:45 +0200 (Thu, 16 Oct 2008) | 3 lines
Added a missing import and spacing fixes.
........
r7777 | bugman | 2008-10-16 18:06:41 +0200 (Thu, 16 Oct 2008) | 3 lines
Another missing model-free multi-processor command.
........
r7778 | bugman | 2008-10-16 18:10:33 +0200 (Thu, 16 Oct 2008) | 3 lines
Fixes for the model-free minimisation setup and execution using the multi-processor structure.
........
r7779 | bugman | 2008-10-16 18:15:06 +0200 (Thu, 16 Oct 2008) | 3 lines
Fixes and clean up of the MF_memo class.
........
r7780 | bugman | 2008-10-16 18:20:06 +0200 (Thu, 16 Oct 2008) | 3 lines
Removed the old, commented out Processor class.
........
r7781 | bugman | 2008-10-16 18:25:38 +0200 (Thu, 16 Oct 2008) | 6 lines
Clean up of the Uni_processor class.
Old, commented out code has been deleted, a few docstrings added, whitespacing completely fixed, and
class methods alphabetically sorted.
........
r7782 | bugman | 2008-10-16 18:29:09 +0200 (Thu, 16 Oct 2008) | 3 lines
Comment fixes.
........
r7783 | bugman | 2008-10-16 18:29:44 +0200 (Thu, 16 Oct 2008) | 3 lines
Docstring fix.
........
r7794 | bugman | 2008-10-17 00:16:35 +0200 (Fri, 17 Oct 2008) | 3 lines
Method spacing fix.
........
r7808 | bugman | 2008-10-18 17:48:26 +0200 (Sat, 18 Oct 2008) | 5 lines
Created a singleton container class called Processor_box for storing the Processors.
This can then be used to get access to the Processor.
........
r7809 | bugman | 2008-10-18 17:55:41 +0200 (Sat, 18 Oct 2008) | 5 lines
When instantiating the Processor via import_module(), all info is now stored in the Processor_box.
The Processor instance, its name, and its class name are all stored in the singleton.
........
r7810 | bugman | 2008-10-18 18:04:36 +0200 (Sat, 18 Oct 2008) | 5 lines
Made the Processor_box inherit from the object object.
This is necessary for the current singleton implementation.
........
r7811 | bugman | 2008-10-18 18:06:25 +0200 (Sat, 18 Oct 2008) | 5 lines
Fixes for the use of the Processor in model-free optimisation.
The Processor_box singleton is instantiated to get access to the Processor instance.
........
r7812 | bugman | 2008-10-18 18:18:56 +0200 (Sat, 18 Oct 2008) | 3 lines
Clean up and fixes for the do_feedback() method.
........
r7813 | bugman | 2008-10-18 18:19:58 +0200 (Sat, 18 Oct 2008) | 3 lines
Renamed print_flag to verbosity to shift to the new 1.3 design.
........
r7814 | bugman | 2008-10-19 15:15:45 +0200 (Sun, 19 Oct 2008) | 3 lines
A bit of a clean up of do_feedback().
........
r7815 | bugman | 2008-10-19 15:19:48 +0200 (Sun, 19 Oct 2008) | 3 lines
Alphabetically ordered the classes where possible.
........
r7816 | bugman | 2008-10-19 15:32:34 +0200 (Sun, 19 Oct 2008) | 3 lines
Big spacing clean ups.
........
r7817 | bugman | 2008-10-19 15:37:45 +0200 (Sun, 19 Oct 2008) | 3 lines
Alphabetical sorting of all class methods.
........
r7818 | bugman | 2008-10-19 15:40:05 +0200 (Sun, 19 Oct 2008) | 3 lines
Removed all references to 'run' and replaced the 'index' integer with the 'spin' SpinContainer.
........
r7819 | bugman | 2008-10-19 15:43:46 +0200 (Sun, 19 Oct 2008) | 3 lines
Renamed 'res_id' to 'spin_id'.
........
r7820 | bugman | 2008-10-19 15:46:59 +0200 (Sun, 19 Oct 2008) | 3 lines
Added 'sim_index' to the info_map dictionary.
........
r7821 | bugman | 2008-10-19 15:48:54 +0200 (Sun, 19 Oct 2008) | 3 lines
Replace the old 'param_set' variable with 'model_type'.
........
r7822 | bugman | 2008-10-19 15:51:38 +0200 (Sun, 19 Oct 2008) | 3 lines
Updated the mf_map MF_minimise_command class instance variable to match the new design.
........
r7823 | bugman | 2008-10-19 16:41:53 +0200 (Sun, 19 Oct 2008) | 6 lines
Removed the minimisation header print out for individual spins.
This was put in specific_fns.model_free.multi_processor_commands.MF_minimise_command.do_feedback()
in Gary's original branch.
........
r7824 | bugman | 2008-10-19 16:50:45 +0200 (Sun, 19 Oct 2008) | 6 lines
Shifted the multiprocessor queue running calls from the prompt interface to the generic_fns code.
This code was in the wrong place as the prompt code is reserved solely for argument checking and
passing these onto the functional code (as well as providing documentation and the user interface).
........
r7825 | bugman | 2008-10-19 17:09:26 +0200 (Sun, 19 Oct 2008) | 3 lines
Commented the new model-free minimisation code.
........
r7826 | bugman | 2008-10-19 17:11:55 +0200 (Sun, 19 Oct 2008) | 3 lines
Fix for the initialisation of the 'b' key in minimise_map. None had quotes around it.
........
r7827 | bugman | 2008-10-19 17:16:33 +0200 (Sun, 19 Oct 2008) | 3 lines
Added some docstrings and comments to MF_minimise_command.
........
r7828 | bugman | 2008-10-19 17:29:03 +0200 (Sun, 19 Oct 2008) | 5 lines
Sent the spin_id and sim_index into set_minimise().
These were missing from the maps!
........
r7829 | bugman | 2008-10-19 17:31:34 +0200 (Sun, 19 Oct 2008) | 3 lines
More comments and docstrings.
........
r7830 | bugman | 2008-10-19 17:33:37 +0200 (Sun, 19 Oct 2008) | 3 lines
Added the missing minfx import and updated the copyright statement.
........
r7831 | bugman | 2008-10-19 17:36:28 +0200 (Sun, 19 Oct 2008) | 3 lines
Fix for the verbosity argument. generic_minimise() requires this arg to be called print_flag.
........
r7832 | bugman | 2008-10-19 17:51:51 +0200 (Sun, 19 Oct 2008) | 3 lines
Clean up of the disassemble_result() method.
........
r7833 | bugman | 2008-10-19 17:52:13 +0200 (Sun, 19 Oct 2008) | 3 lines
Removed some unused code.
........
r7834 | bugman | 2008-10-19 18:05:08 +0200 (Sun, 19 Oct 2008) | 3 lines
Put some spacing back.
........
r7835 | bugman | 2008-10-19 18:09:26 +0200 (Sun, 19 Oct 2008) | 3 lines
Removed an unused import (and incorrect import).
........
r7836 | bugman | 2008-10-19 18:20:05 +0200 (Sun, 19 Oct 2008) | 3 lines
Removed the svn:executable property from the 2 processor modules.
........
r7838 | bugman | 2008-10-19 19:05:02 +0200 (Sun, 19 Oct 2008) | 3 lines
Bug fix for the arguments() checking when not in uni-processor mode.
........
r7839 | bugman | 2008-10-19 19:10:52 +0200 (Sun, 19 Oct 2008) | 3 lines
Added error checking for the processor type passed in with the -m command line argument.
........
r7840 | bugman | 2008-10-19 20:16:17 +0200 (Sun, 19 Oct 2008) | 3 lines
Fixed the checks for the supported processor fabric types.
........
r7841 | bugman | 2008-10-19 20:17:48 +0200 (Sun, 19 Oct 2008) | 3 lines
Spelling fix for the raise_unimplemented() function.
........
r7842 | bugman | 2008-10-19 21:25:23 +0200 (Sun, 19 Oct 2008) | 6 lines
Massive clean up of the multi.multi_processor module.
Large amounts of whitespace inconsistencies have been cleaned up, the classes and their methods
alphabetically sorted, the imports sorted out, a number of docstrings and comments added, etc.
........
r7843 | bugman | 2008-10-19 21:31:45 +0200 (Sun, 19 Oct 2008) | 6 lines
Renamed the multi_processor module to multi_processor_base.
This was mentioned in the module and by Gary in the detailed post:
https://mail.gna.org/public/relax-devel/2007-05/msg00000.html.
........
r7844 | bugman | 2008-10-19 21:33:23 +0200 (Sun, 19 Oct 2008) | 3 lines
Fixed the import of multi_processor_base.
........
r7845 | bugman | 2008-10-19 21:36:13 +0200 (Sun, 19 Oct 2008) | 3 lines
Shifted a class out of order to allow for subclassing.
........
r7846 | bugman | 2008-10-19 21:42:43 +0200 (Sun, 19 Oct 2008) | 6 lines
Clean up of the mpi4py multi-processor module.
This includes the copyright notice, whitespace, and imports. One import has been fixed to use the
new multi_processor_base module.
........
r7847 | bugman | 2008-10-19 21:44:44 +0200 (Sun, 19 Oct 2008) | 3 lines
Alphabetical ordering of functions and methods.
........
r7848 | bugman | 2008-10-19 21:52:49 +0200 (Sun, 19 Oct 2008) | 3 lines
Bug fix. MPI.size and MPI.rank are now located in MPI.COMM_WORLD!
........
r7849 | bugman | 2008-10-19 21:54:52 +0200 (Sun, 19 Oct 2008) | 3 lines
Some more fixes for MPI.COMM_WORLD.size and MPI.COMM_WORLD.rank.
........
r7850 | bugman | 2008-10-19 23:12:34 +0200 (Sun, 19 Oct 2008) | 5 lines
Updates to the get_stdio_pre_strings() method including replacing ">" with "|".
The change is for better segregation of the pre-strings and the relax print outs.
........
r7851 | bugman | 2008-10-19 23:24:33 +0200 (Sun, 19 Oct 2008) | 6 lines
More playing around with the STDOUT and STDERR prepending.
The 'S' character has been removed from the STDOUT stream. The 'E' error character in the STDERR
stream is now much more obvious. This could be replaced or restored with discussions.
........
r7853 | bugman | 2008-10-20 20:54:49 +0200 (Mon, 20 Oct 2008) | 3 lines
Spacing fixes and a few added comments.
........
r7854 | bugman | 2008-10-20 21:20:55 +0200 (Mon, 20 Oct 2008) | 3 lines
Added a newline character to the end of the file.
........
r7855 | bugman | 2008-10-20 21:21:44 +0200 (Mon, 20 Oct 2008) | 3 lines
Added the copyright statement.
........
r7856 | bugman | 2008-10-20 21:25:25 +0200 (Mon, 20 Oct 2008) | 3 lines
A large number of spacing fixes.
........
r7857 | bugman | 2008-10-20 21:27:28 +0200 (Mon, 20 Oct 2008) | 3 lines
Alphabetical sorting of class methods.
........
r7858 | bugman | 2008-10-20 21:28:24 +0200 (Mon, 20 Oct 2008) | 3 lines
Removed a patch of debugging code.
........
r7859 | bugman | 2008-10-20 21:42:54 +0200 (Mon, 20 Oct 2008) | 3 lines
Shortened the copyright notices.
........
r7860 | bugman | 2008-10-20 21:44:12 +0200 (Mon, 20 Oct 2008) | 3 lines
Bug fix - removed a trailing ',' character.
........
r7861 | bugman | 2008-10-20 21:45:39 +0200 (Mon, 20 Oct 2008) | 3 lines
Whitespace fix.
........
r7862 | bugman | 2008-10-20 21:48:35 +0200 (Mon, 20 Oct 2008) | 3 lines
Alphabetical method sorting.
........
r7863 | bugman | 2008-10-20 21:56:41 +0200 (Mon, 20 Oct 2008) | 5 lines
Modified the get_intro_string() methods.
Epydoc formatted docstrings have been added and the text slightly modified.
........
r7864 | bugman | 2008-10-20 21:58:29 +0200 (Mon, 20 Oct 2008) | 3 lines
Turned off the "Processor model:" printout when in prompt mode with the uni-processor fabric.
........
r7865 | bugman | 2008-10-20 22:04:45 +0200 (Mon, 20 Oct 2008) | 3 lines
Modified the "overall runtime" print out (start with \n and capitalised).
........
r7875 | bugman | 2008-10-20 23:40:01 +0200 (Mon, 20 Oct 2008) | 3 lines
The test suite and licence mode of relax are now allowed in multi-processor mode.
........
r9408 | bugman | 2009-08-28 11:37:48 +0200 (Fri, 28 Aug 2009) | 3 lines
Docstring fix.
........
r9448 | bugman | 2009-09-03 14:53:02 +0200 (Thu, 03 Sep 2009) | 3 lines
Added a unit test directory for the multi package code.
........
r9449 | bugman | 2009-09-03 14:56:00 +0200 (Thu, 03 Sep 2009) | 3 lines
Added a unit test for the multi package __all__ list.
........
r9450 | bugman | 2009-09-03 14:57:23 +0200 (Thu, 03 Sep 2009) | 3 lines
Fix for the multi package __all__ list.
........
r9451 | bugman | 2009-09-03 15:03:52 +0200 (Thu, 03 Sep 2009) | 6 lines
Manually changed a call to type() to isinstance().
This was because the following command failed:
2to3 -f types ./multi/processor.py
........
r9452 | bugman | 2009-09-03 15:13:50 +0200 (Thu, 03 Sep 2009) | 17 lines
Updates for Python3k.
The command used was:
2to3 -w \
-x except \
-x import \
-x imports \
-x long \
-x numliterals \
-x xrange \
relax .
This is from the docs/2to3_checklist document. Two print statements changed in multi/processor.py
had to be manually fixed afterwards, as the end and file args to the print statement do not exist in
all python 2.x versions.
........
r9453 | bugman | 2009-09-03 15:20:11 +0200 (Thu, 03 Sep 2009) | 19 lines
Updates for Python3k.
The command used was:
2to3 -w \
-f buffer \
-f idioms \
-f set_literal \
-f ws_comma \
-x except \
-x import \
-x imports \
-x long \
-x numliterals \
-x xrange \
relax .
This is from the docs/2to3_checklist document.
........
r9997 | bugman | 2009-11-27 14:57:08 +0100 (Fri, 27 Nov 2009) | 5 lines
Fixed a bug introduced in r9996.
The mf_minimise module conflicts were not properly resolved. This should now work better.
........
r10119 | bugman | 2010-01-06 19:12:32 +0100 (Wed, 06 Jan 2010) | 5 lines
Bug fix for IO redirection!
This may cause problems, but now IO redirection has been fixed.
........
r10120 | bugman | 2010-01-06 19:19:52 +0100 (Wed, 06 Jan 2010) | 6 lines
Reverted r10119 as this was not fixing the IO redirection problem.
The command used was:
svn merge -r10119:10118 .
........
r10121 | bugman | 2010-01-06 19:20:59 +0100 (Wed, 06 Jan 2010) | 5 lines
Modified the IO redirection methods in the Processor base class.
The original IO streams are stored and then restored instead of sys.__stdout__ and sys.__stderr__.
........
r10122 | bugman | 2010-01-07 11:41:06 +0100 (Thu, 07 Jan 2010) | 5 lines
Removed all STDOUT and STDERR handling in the multi-processor model-free specific code.
This is 100% handled by the multi module.
........
r10123 | bugman | 2010-01-07 11:43:51 +0100 (Thu, 07 Jan 2010) | 6 lines
Bug fix for the STDOUT and STDERR handling by the multi module.
The IO streams are no longer lost in the case of an error. This allows the relax test suite to
function correctly in uni-processor mode. The multi-processor mode might require more work.
........
r10124 | bugman | 2010-01-07 13:49:17 +0100 (Thu, 07 Jan 2010) | 5 lines
Fix for the run_command_queue() method.
This was broken by the 2to3 command!
........
r10125 | bugman | 2010-01-07 13:50:25 +0100 (Thu, 07 Jan 2010) | 3 lines
Second attempt at r10124. The last commit contained no code changes!
........
r10126 | bugman | 2010-01-07 14:38:22 +0100 (Thu, 07 Jan 2010) | 3 lines
Updated of the copyright statement and removal of dead imports.
........
r10127 | bugman | 2010-01-07 14:38:59 +0100 (Thu, 07 Jan 2010) | 3 lines
Bug fix, the disassemble_result() method is now out of the specific API.
........
r10128 | bugman | 2010-01-07 14:40:30 +0100 (Thu, 07 Jan 2010) | 6 lines
Fixes for the process_results() method.
The result_string variable was accidentally deleted in the attempt to fix the STDOUT and STDERR
problems.
........
r10129 | bugman | 2010-01-07 14:40:58 +0100 (Thu, 07 Jan 2010) | 3 lines
Removed a ton of dead code. This was all commented out.
........
r10130 | bugman | 2010-01-07 14:43:55 +0100 (Thu, 07 Jan 2010) | 3 lines
Trailing whitespace fixes.
........
r10131 | bugman | 2010-01-07 15:55:07 +0100 (Thu, 07 Jan 2010) | 3 lines
The ability to perform code profiling has been restored in the multi_processor branch!
........
r10132 | bugman | 2010-01-07 15:55:54 +0100 (Thu, 07 Jan 2010) | 3 lines
Removed an unneeded FIXME statement.
........
r10133 | bugman | 2010-01-07 15:58:38 +0100 (Thu, 07 Jan 2010) | 3 lines
Some clean ups of the model-free code for multi-processing.
........
r10134 | bugman | 2010-01-07 16:00:08 +0100 (Thu, 07 Jan 2010) | 3 lines
Profiling was accidentally turned on!
........
r10135 | bugman | 2010-01-07 16:51:15 +0100 (Thu, 07 Jan 2010) | 3 lines
Added more sanity to the STDOUT and STDERR handling by adding some flushing.
........
r10136 | bugman | 2010-01-07 16:56:05 +0100 (Thu, 07 Jan 2010) | 3 lines
Removed an unused import.
........
r10137 | bugman | 2010-01-07 17:00:40 +0100 (Thu, 07 Jan 2010) | 6 lines
Code simplification.
The build_mf() and run_command() methods have been collapsed into the run() method, significantly
cutting the amount of jumping around in the code.
........
r10138 | bugman | 2010-01-07 17:03:16 +0100 (Thu, 07 Jan 2010) | 3 lines
Removed the unused pre_run() method.
........
r10139 | bugman | 2010-01-07 17:07:30 +0100 (Thu, 07 Jan 2010) | 3 lines
Added some docstrings and comments.
........
r10140 | bugman | 2010-01-07 17:08:21 +0100 (Thu, 07 Jan 2010) | 3 lines
Deleted the useless post_run() method.
........
r10141 | bugman | 2010-01-07 17:33:05 +0100 (Thu, 07 Jan 2010) | 9 lines
Clean up of the code before and after minfx execution.
This includes a number of simplifications. pre_command_feed_back() is now called
initial_printout(). post_command_feedback() has been deleted. This was for a kludge which can be
avoided by setting the minfx printout arg to zero. The calls to set_generic_pre_and_post_amble()
and set_grid_pre_and_post_amble() were part of this kludge and are eliminated. The self.silence
flag is now used to silence printouts when desired.
........
r10142 | bugman | 2010-01-07 17:35:17 +0100 (Thu, 07 Jan 2010) | 3 lines
Whitespace fix.
........
r10143 | bugman | 2010-01-07 17:36:46 +0100 (Thu, 07 Jan 2010) | 3 lines
Import fixes. Minfx is only used in multi_processor_commands!
........
r10144 | bugman | 2010-01-07 17:59:00 +0100 (Thu, 07 Jan 2010) | 7 lines
Simplification of the model-free target function class initialisation.
This has been shifted back into specific_fns/model_free/mf_minimise.py! This does not need to be
done by the multi-processing code. Silencing can be performed later by modifying the 'verbosity'
variable in this class.
........
r10145 | bugman | 2010-01-07 18:00:59 +0100 (Thu, 07 Jan 2010) | 3 lines
Modified initial_printout() to match the main 1.3 line code.
........
r10146 | bugman | 2010-01-07 18:03:40 +0100 (Thu, 07 Jan 2010) | 3 lines
Added a run() method to the MF_grid_command class and removed the dummy initial_printout() method.
........
r10147 | bugman | 2010-01-07 18:04:46 +0100 (Thu, 07 Jan 2010) | 3 lines
Absorbed initial_printout() into run() in the MF_minimise_command class.
........
r10148 | bugman | 2010-01-07 18:07:48 +0100 (Thu, 07 Jan 2010) | 5 lines
Eliminated the silence flag in all classes.
This can be reinstated later, if necessary, solely by the diffusion tensor grid search.
........
r10149 | bugman | 2010-01-07 18:08:28 +0100 (Thu, 07 Jan 2010) | 3 lines
Removed the unneeded 2nd initialisation of the Mf class during back-calculation.
........
r10150 | bugman | 2010-01-07 18:09:33 +0100 (Thu, 07 Jan 2010) | 3 lines
Comment removal to match the main 1.3 line.
........
r10151 | bugman | 2010-01-07 18:18:47 +0100 (Thu, 07 Jan 2010) | 5 lines
The normal grid search has been activated using MF_grid_command.
The diffusion tensor sub-grid algorithm might need a MF_subgrid_command class.
........
r10152 | bugman | 2010-01-07 18:44:20 +0100 (Thu, 07 Jan 2010) | 7 lines
A number of changes for the MF_grid_command class.
process_results() has been absorbed into run(), the data required for the grid search is passed into
the class upon initialisation, and the call to minfx.generic_minimise() has been switched to a call
to minfx.grid() to correspond to the minfx library changes.
........
r10153 | bugman | 2010-01-07 18:45:32 +0100 (Thu, 07 Jan 2010) | 3 lines
Absorbed process_results() into run() in the MF_minimise_command class
........
r10154 | bugman | 2010-01-07 18:58:09 +0100 (Thu, 07 Jan 2010) | 6 lines
Eliminated the MF_minimise_command.set_minimise() method.
All data is now sent into the __init__() method. The MF_grid_command class is now subclassed from
Slave_command to avoid the MF_minimise_command.__init__() method.
........
r10155 | bugman | 2010-01-08 10:07:11 +0100 (Fri, 08 Jan 2010) | 3 lines
Removed a duplicated print out. This printout is in multi_processor_commands.
........
r10156 | bugman | 2010-01-08 10:24:13 +0100 (Fri, 08 Jan 2010) | 3 lines
Improved the printouts for model-free optimisation by adding the spin_print() function.
........
r10159 | bugman | 2010-01-11 10:45:18 +0100 (Mon, 11 Jan 2010) | 3 lines
Added some docstrings.
........
r10160 | bugman | 2010-01-11 10:45:50 +0100 (Mon, 11 Jan 2010) | 3 lines
Added the dummy post_run() method to prevent the overall time printout in uni-processor mode.
........
r10161 | bugman | 2010-01-11 10:50:46 +0100 (Mon, 11 Jan 2010) | 3 lines
Added a copyright statement.
........
r10162 | bugman | 2010-01-11 10:51:43 +0100 (Mon, 11 Jan 2010) | 5 lines
Ported the multi.mpi4py_processor module to the new mpi4py interface.
The sending and receiving of objects is now performed differently.
........
r10163 | bugman | 2010-01-11 10:52:36 +0100 (Mon, 11 Jan 2010) | 3 lines
Shifted the mpi4py dependency check into the dep_check module.
........
r10164 | bugman | 2010-01-11 10:53:14 +0100 (Mon, 11 Jan 2010) | 3 lines
The mpi4py dep check error is now performed by the base relax module.
........
r10165 | bugman | 2010-01-11 10:55:19 +0100 (Mon, 11 Jan 2010) | 6 lines
Changes to how the model-free and optimisation data/params are passed into the processor instances.
Two data containers are now used, one for the model-free data, the other for the optimisation
parameters. This significantly cleans up the multi-processing model-free interface.
........
r10166 | bugman | 2010-01-11 11:10:14 +0100 (Mon, 11 Jan 2010) | 9 lines
Copied the grid.py minimisation module from the original multi_processor branch.
The command used was:
svn cp svn+ssh://bugman@.../svn/relax/branches/multi_processor/minimise/grid.py subdivide_grid.py
This module is now part of minfx, but the Grid_info class will be used by the multi-processor code to
subdivide the diffusion tensor grid search. The code needs much work!
........
r10167 | bugman | 2010-01-11 11:18:08 +0100 (Mon, 11 Jan 2010) | 3 lines
Many, many whitespace fixes.
........
r10168 | bugman | 2010-01-11 11:22:21 +0100 (Mon, 11 Jan 2010) | 3 lines
Conversion from Numeric to numpy.
........
r10169 | bugman | 2010-01-11 11:23:07 +0100 (Mon, 11 Jan 2010) | 3 lines
Eliminated the pre_and_post_amble mess (that can be nicely handled with the verbosity flag).
........
r10170 | bugman | 2010-01-11 11:24:36 +0100 (Mon, 11 Jan 2010) | 3 lines
The Grid_info class is now being imported.
........
r10171 | bugman | 2010-01-11 13:00:33 +0100 (Mon, 11 Jan 2010) | 5 lines
Modified the Grid_info class remove the grid_ops arg.
This is now removed and subdivided into 'inc', 'lower', and 'upper' in the 1.3 line.
........
r10172 | bugman | 2010-01-11 17:06:53 +0100 (Mon, 11 Jan 2010) | 8 lines
Redesigned and cleaned up how the Grid_info class is initialised.
This is necessary as the grid_ops data structure was eliminated from the 1.3 line a long time ago.
Now the lower, upper, and inc data structure structures are passed in. In __init__(),
calc_grid_size(), calc_grid_values(), and calc_strides(), the code has been cleaned up, comments
added, and docstrings written.
........
r10173 | bugman | 2010-01-11 17:10:01 +0100 (Mon, 11 Jan 2010) | 3 lines
Alphabetical ordering of methods.
........
r10174 | bugman | 2010-01-11 17:11:56 +0100 (Mon, 11 Jan 2010) | 3 lines
Removed the nesting of the Iterator class.
........
r10175 | bugman | 2010-01-11 18:14:12 +0100 (Mon, 11 Jan 2010) | 3 lines
Added a docstring to the __iter__() method.
........
r10176 | bugman | 2010-01-11 18:19:01 +0100 (Mon, 11 Jan 2010) | 3 lines
Converted a number of methods to private methods.
........
r10177 | bugman | 2010-01-11 19:14:10 +0100 (Mon, 11 Jan 2010) | 6 lines
Clean up of the sub_divide() method.
A docstring has been written, the sub_grid() method has been absorbed into one line of sub_divide()
simplifying the code paths, and comments have been added coving all statements.
........
r10189 | bugman | 2010-01-13 11:12:49 +0100 (Wed, 13 Jan 2010) | 3 lines
Added comments and docstrings to the MF_memo and MF_result_command classes.
........
r10190 | bugman | 2010-01-13 11:16:01 +0100 (Wed, 13 Jan 2010) | 3 lines
Argument rearrangement.
........
r10191 | bugman | 2010-01-13 11:22:05 +0100 (Wed, 13 Jan 2010) | 7 lines
Bug fix for the model-free optimisation.
The data containers used to store all the model-free info and optimisation parameters was being
overwritten so each command sent to the processor queue contained only the data of the last spin
system.
........
r10192 | bugman | 2010-01-13 11:36:48 +0100 (Wed, 13 Jan 2010) | 5 lines
Cleaned up the run_command_queue() method idle and running set print outs.
The code has also been commented.
........
r10193 | bugman | 2010-01-13 11:37:30 +0100 (Wed, 13 Jan 2010) | 3 lines
Removed the unnecessary IO buffer flush added in the last commit.
........
r10194 | bugman | 2010-01-13 11:37:51 +0100 (Wed, 13 Jan 2010) | 3 lines
Comment simplifications.
........
r10195 | bugman | 2010-01-13 11:47:35 +0100 (Wed, 13 Jan 2010) | 3 lines
Added more comments to run_command_queue().
........
r10196 | bugman | 2010-01-13 14:06:11 +0100 (Wed, 13 Jan 2010) | 5 lines
Removed unused code from MF_result_command.run().
The minimisation stats don't need to be placed into the memo!
........
r10198 | bugman | 2010-01-13 15:41:04 +0100 (Wed, 13 Jan 2010) | 3 lines
Expanded the run() docstring to explain why this method exists.
........
r10199 | bugman | 2010-01-13 15:48:10 +0100 (Wed, 13 Jan 2010) | 3 lines
Fixes for the examples given in the multi.processor module docstring.
........
r10200 | bugman | 2010-01-13 15:49:34 +0100 (Wed, 13 Jan 2010) | 3 lines
More module docstring example fixes.
........
r10201 | bugman | 2010-01-13 15:56:12 +0100 (Wed, 13 Jan 2010) | 3 lines
Added the subdivide_grid module to the __all__ package list.
........
r10202 | bugman | 2010-01-13 16:05:04 +0100 (Wed, 13 Jan 2010) | 7 lines
Broke some nasty recursion.
The processor instance was being placed into the relax instance and then relax calling the processor
instance method get_intro_string() to get a message about the processor. Now the top level code is
calling the method and placing the string into the relax instance.
........
r10203 | bugman | 2010-01-13 16:07:53 +0100 (Wed, 13 Jan 2010) | 3 lines
Renamed relax_instance to simply relax.
........
r10758 | bugman | 2010-02-18 19:03:19 +0100 (Thu, 18 Feb 2010) | 5 lines
Preparations for restoring Gary's model-free diffusion tensor grid search parallelisations.
The MF_split_grid_command will be the new name for this old code.
........
r10760 | bugman | 2010-02-18 19:11:16 +0100 (Thu, 18 Feb 2010) | 3 lines
The get_intro_string() call is no longer valid, it was removed from the 1.3 line.
........
r10761 | bugman | 2010-02-18 23:29:31 +0100 (Thu, 18 Feb 2010) | 5 lines
Renamed MF_super_grid_memo to MF_split_grid_memo.
This is the memo for the parallelised grid!
........
r10762 | bugman | 2010-02-18 23:37:36 +0100 (Thu, 18 Feb 2010) | 6 lines
Only the minimal amount of info is being passed into MF_split_grid_memo.
The other info can be passed in through a command.store_data() call. This mimics the current
MF_memo operation.
........
r10763 | bugman | 2010-02-18 23:50:56 +0100 (Thu, 18 Feb 2010) | 6 lines
Clean up of the MF_split_grid_memo.__init__() method.
This now matches MF_memo, but with a few additional variable initialisations (not all may be needed
in the end).
........
r10764 | bugman | 2010-02-18 23:53:35 +0100 (Thu, 18 Feb 2010) | 5 lines
The MF_split_grid_command.store_data() method is being used to pump in all the data.
This comes from the base class.
........
r10765 | bugman | 2010-02-19 00:00:55 +0100 (Fri, 19 Feb 2010) | 5 lines
MF_grid_memo is being used to for the parallel grid sub-elements.
This was how the original code worked, not sure why this changed?
........
r10790 | bugman | 2010-02-21 19:32:12 +0100 (Sun, 21 Feb 2010) | 6 lines
Switched to using the same classes as normal multi-model-free for the diffusion tensor grid search.
The model-free code setup is far, far too complex. The diffusion tensor parallel grid search
classes will be eliminated, and the normal model-free multi-command classes used for this purpose.
........
r10791 | bugman | 2010-02-21 19:37:31 +0100 (Sun, 21 Feb 2010) | 9 lines
Deletion of the parallelised grid search classes.
The model-free multi-processor code is now far simpler. However the parallelised grid search is not
functioning correctly yet. The deleted classes are:
MF_grid_memo
MF_grid_result_command
MF_split_grid_command
........
r10792 | bugman | 2010-02-21 19:40:34 +0100 (Sun, 21 Feb 2010) | 3 lines
Removed the dead multi_processor_commands imports and deleted the super_grid_memo.
........
r10793 | bugman | 2010-02-21 20:11:37 +0100 (Sun, 21 Feb 2010) | 3 lines
Deleted the MF_split_grid_memo class. This should have been deleted in a previous commit!
........
r10830 | bugman | 2010-02-22 22:22:31 +0100 (Mon, 22 Feb 2010) | 5 lines
Switched from using Grid_info to minfx.grid.grid_split().
The model-free grid search 'command' will need to be modified to utilise this.
........
r10831 | bugman | 2010-02-22 22:42:39 +0100 (Mon, 22 Feb 2010) | 3 lines
The subdivided grid search is now activated in the MF_grid_command class.
........
r10832 | bugman | 2010-02-23 00:33:31 +0100 (Tue, 23 Feb 2010) | 5 lines
Modified _disassemble_result() to be parallelisation friendly.
Optimisation results are only stored if the chi2 value is lower!
........
r10833 | bugman | 2010-02-23 00:39:53 +0100 (Tue, 23 Feb 2010) | 3 lines
Fix for the _disassemble_result() method for when there is no chi2 variable present.
........
r10834 | bugman | 2010-02-23 01:39:14 +0100 (Tue, 23 Feb 2010) | 6 lines
Fixes and simplification of the PrependOut class.
Much code and comments have been deleted to significantly simplify this class. The stream is now
flushed with a flush() call. This fixes the problem of STDERR and STDOUT interleaving!
........
r10835 | bugman | 2010-02-23 02:26:46 +0100 (Tue, 23 Feb 2010) | 3 lines
PrependOut.write() now flushes both STDOUT and STDERR giving perfect interleaving of the 2 streams.
........
r10836 | bugman | 2010-02-23 02:27:58 +0100 (Tue, 23 Feb 2010) | 6 lines
Deleted the generic_fns.subdivide_grid module.
The multi-processing code is fully functional without it as the functionality has been shifted to
minfx.
........
r10837 | bugman | 2010-02-23 02:34:49 +0100 (Tue, 23 Feb 2010) | 3 lines
Removed 'subdivide_grid' from the __all__ list.
........
r10840 | bugman | 2010-02-23 02:48:43 +0100 (Tue, 23 Feb 2010) | 5 lines
Fix for the relax run() method.
The previous merge needed to be modified to function with the changes to the relax base module.
........
r10844 | bugman | 2010-02-23 03:02:53 +0100 (Tue, 23 Feb 2010) | 3 lines
Another fix for the 'mode' variable being located in the class namespace.
........
r10861 | bugman | 2010-02-23 17:36:09 +0100 (Tue, 23 Feb 2010) | 7 lines
Fix for the Capturing_exception class.
Sometimes the exc_info does not correspond to an exception (i.e. sys.exc_info() returns None). This
is now caught and the class returns silently. The result is that the true traceback from the slaves
is printed!
........
r10864 | bugman | 2010-02-23 18:29:59 +0100 (Tue, 23 Feb 2010) | 3 lines
The processor fabric intro string is now placed into the info box.
........
r10867 | bugman | 2010-02-23 18:39:14 +0100 (Tue, 23 Feb 2010) | 3 lines
Fixes for the name change of the mpi4py module import flag.
........
r10868 | bugman | 2010-02-23 18:41:16 +0100 (Tue, 23 Feb 2010) | 3 lines
Reintroduced into the intro string the processor fabric info print out.
........
r10869 | bugman | 2010-02-23 18:56:33 +0100 (Tue, 23 Feb 2010) | 3 lines
The multi-processor intro string is now word wrapped to the given console width.
........
r10870 | bugman | 2010-02-23 18:57:38 +0100 (Tue, 23 Feb 2010) | 5 lines
Added vendor info to the mpi4py processor intro string
I.e. Open MPI 1.3.3 is now printed out as well to help with debugging.
........
r12420 | bugman | 2011-01-21 16:08:28 +0100 (Fri, 21 Jan 2011) | 5 lines
Shifted the relax warning system setup into the run() method as it can no longer be in __init__().
It requires the command line args, but these are processed after class initialisation in this branch.
........
r12421 | bugman | 2011-01-21 16:11:01 +0100 (Fri, 21 Jan 2011) | 3 lines
Fix for the run() method - the relax operation mode is stored in 'self.mode' in this branch.
........
................
r14388 | bugman | 2011-08-16 13:52:09 +0200 (Tue, 16 Aug 2011) | 3 lines
Removed merge tracking for "svnmerge" for
svn+ssh://bugman@.../svn/relax/branches/multi_processor_merge
................
r14390 | bugman | 2011-08-16 18:59:34 +0200 (Tue, 16 Aug 2011) | 3 lines
Removed the unused traceit() function.
................
r14391 | bugman | 2011-08-16 19:03:36 +0200 (Tue, 16 Aug 2011) | 3 lines
Errors in the multi-processor slaves are now sent to sys.stderr.
................
bugman 2011-08-16