1 This version of PsN is tested with perl-5.8.4
2 Depending on your distribution you may have to install the following
8 All available from CPAN ( www.cpan.org )
11 ----------------------------------------------
13 Changes from 2.2.0 to 2.3.0
16 - Dropped use of parallell forkmanager.
18 - Failed runs are now ignored, they are indicated as "failed
19 run" in "raw_results". (And a warning will be printed).
21 - The default debug level has been raised from 0(only fatal messages
22 will be printed) to 1 (fatal and warning messages). The biggest
23 benefit is that a lot of messages can be turned of by setting debug
24 to 0. Also it will force us, the developers, to rethink the warning
25 messages a bit, making them more informative.
27 - The config for diffrent nonmem versions now requires the actual
28 nonmem version to be specified. Like this:
31 name=/path/to/nonmem,5
32 name_six=/path/to/nonmem6,6
34 - PsN will occupy slightly less disc space when running.
37 - If someone specified a maxeval bigger than 9999 PsN would try and
38 fix that, and failed. Fixed.
40 Changes from 2.1.10 to 2.2.0 ( released for PsN course )
44 - PsN Output module use a proper Matrix library for handling some of
45 the matrices of the outputfile.
47 - The bootstrap can now be resumed with a lower number of
48 samples. Meaning that if you are in a hurry and want to stop the
49 bootstrap and get results after 1000 out of 2000 runs, you can. At
50 a later time you can then resume the 2000 runs.
52 - Nonmem warnings are now printed if "--verbose" is givven.
54 - Automatic incrementation of directory names are a bit smarter (no
55 more "directory modelfit_dirX is in the way" messages).
57 - Clean is much more stringent and intelligent and now manages to
58 remove entire tool directories if desired.
60 - option list printed from "-h|-?" only shows utility specific options.
62 - Default values configurable on a per utility basis.
64 - Online documentation improved
66 - Parallel run of NONMEM is not aborted as before if a outputfile is
67 not correctly parsed. It is just ignored, as if it had not
68 minimized successfully. (actually it is the output parser that
69 doesn't die, but the effect is the same)
71 - Parallel run is a bit nicer to the system when compiling many
72 NONMEMS. By adding a configurable delay between starts.
74 - Matlab and R scripts for various plots from cdd and bootstrap
75 results are distributed with PsN.
77 - Vast improvments in results presentations from all tools.
79 - Improvements in rerun algoritm with new "min_retries" option.
81 - Output from running utilties are much nicer.
83 - new "-quick_summary" prints minimization message and ofv value for
84 each finished run. As does "-verbose" option which also prints
85 "R:X" for each retrie (X is the retrie number).
87 - new "-summary" options prints diagnostics about the tool run when
88 it is finished. For "execute" the "-summary" option will have the
89 same function as the "sumo" utiltiy as well as some diagnostics.
91 - PsN now recognizes that control files may require fortran files and
92 copies them to the run directory. No need to add fotran files to
93 the --extra_files option any more.
95 - Added --extra_output which allows you to copy more than the output
96 file from the run directory.
100 - PsN couldn't parse IGNORE=(ITEM=VALUE) options in $DATA. Fixed :)
102 - If star notation was used in the scm config file section [inits],
103 [upper_bounds] or [lower_bounds] there was a risk of defining
104 relations that was not specified in [test_relations]. Which in turn
105 would cause a lot of diffrent errors. Fixed :)
107 - Using low_cutoff and hi_cutoff in data::sd resulted in PsN trying to use
108 the new compact data strings as the old arrays. fixed.
110 - Removed unecessary regular expression in output file parser.
112 - Not really a bug. But i tried to add "FIX" as a record option to a theta
113 records. Then when "store_init" was called on the "FIX" option PsN crashed
114 because the option was not a "theta_option". So i added a check for
115 "store_init" abilties... if FIX would become a record option in the future.
117 - Member accessor "datafile" replaced with "datafiles" in model. Affected
118 only the bootstrap which have been updated.
120 - An assumptions about outputfiles crashed sumo occationally. Some more
121 rigorous contentchecking has been added and fixed the problem.
124 - Cleanup of dubious utility options.
125 - Option help sorted alphabetically.
126 - "flush", "target" removed from model.
127 - "remove_temp_files" removed from modelfit.
128 - Changed the Makefile considerably, much simpler now.
129 - added "column_to_array" function in data module.
131 Changes from 2.1.9 to 2.1.10
134 - The 'picky' option of all utilities now triggers on
135 COVARIANCE STEP ABORTED and
136 PROGRAM TERMINATED BY OBJ
138 Changes from 2.1.8 to 2.1.9
141 - Nordugrid support reintroduced (but will probably not be developed further)
142 - Sumo now supports multiple problems and subproblems.
143 - Sumo has new option "precision" which control the number of digits
144 to present estimates with.
145 - An extra 'short' logfile is output from the scm with the most important
149 - Bootstrap option "stratify_on" doesn't work with columns defined in
150 extra data files given with "extra_data" option. Not Fixed.
151 - Resumes of scm's with extra_data_files is now possible. Still some
152 problems with scm's where search_direction=both.
153 - LSF options was not communicated properly between tools and subtools
154 so submission of bootstraps, llp, cdd and scm did not work. Fixed.
155 - Default values for LSF can now be configured in psn.conf.
156 - Print order of modelfiles places $MIX before $PRED. Fixed.
157 - extra_files option was broken. Fixed.
160 - Tool will always fork subtools, this will save memory in the scm (and
161 other recursive tools).
162 - Common options need only be specified once. In one of three categories:
163 model_options, tool_options, script_options.
164 - models can now be created with "reference" objects. This solved the
166 - Added function to model that cleans extra data code.
168 Changes from 2.1.7 to 2.1.8
171 - Rudimentary support for execution on Platform LSF grid systems is now
172 included. The following extra attributes are available to all tools
173 for controlling the execution:
176 lsf_queue - Use this queue on the grid
177 lsf_options - General options for the grid queuing system.
178 lsf_project_name - Optional. May be used by the grid system
179 administrator to register control project
181 lsf_job_name - Optional. May be used by a grid system
182 administrator to control multiple runs on a
185 - All tools now have a raw_results file holding the diagnotics and
186 parameter estimates for each NONMEM run.
187 - All tools now handles model files containing one $PROBLEM and no
189 - The Bootstrap and CDD routines now create separate files holding
190 information about which individuals that were included/excluded
191 in/from each created data set.
192 - The Bootstrap results file now has warnings section to better
193 display which runs that were used for result calculations and which
195 - PsN now tries to read psn.conf from the users home directory. (Will
196 require some extra documentation).
197 - Optimized modelfit thread operation by removing Storable::Store.
198 - The clean options has four levels of operation.
200 0 - No cleaning at all.
201 1 - Clean more, but make sure resumes work (this is default).
202 2 - Clean more, but make sure summarize work.
203 3 - Clean everything, keeping only results and logs.
205 - Some attributes and accessors of the PsN-Core class 'output' have
206 had their names changed to make it more clear what their purposes
208 termination is now called minimization_successful
209 termination_message is now called minimization_message
210 covstep_termination is now called covariance_step_successful
212 these attributes are new:
214 covariance_step_warnings (boolean; 0 or 1)
215 estimate_near_boundary (boolean; 0 or 1)
217 The attribute which was previously called termination could have the
218 values 0, 0.5 and 1 for 'minimization terminates', 'minimization
219 successful but warnings or errors were printed from the covariance
220 step' and finally 'minimization successful'. The new attribute
221 'minimization_successful' can only take the values 0 and 1 and the
222 new attribute covariance_step_warnings now takes the value 1 when
223 covariance step warnings were found. Checks are also perfomed to
224 test whether the estimates are close to any boundary.
225 - Html documents is now generated from command line help text.
226 which ensures that they are consistent and up to date.
227 - Personal configuration file now available under Unix like OS:s
230 - Options specific to modelfiles were ignored in some Utilties.
231 - Gradients now output in windows. Just like with nmfe.
232 - Bootstrap now consumes considerably less memory.
233 - Removed unused/obsolete "resume" function from modelfit.
234 - Sorted method entries in modelfit diagram.
235 - Fixed bugs in parsing of OFV and eigenvalues.
237 Changes from 2.1.6 to 2.1.7
240 - "clean" and "compress" options now work under Windows.
241 - SCM now handles crashed runs appropriately.
242 - The ofv_forward, ofv_backward, p_forward and p_backward options of
243 scm now works as they should.
246 - SCM improvements. The final model is now available in the
247 first-level scm directory of each step.
248 - The Utilties -help option is improved. You can now give a list of
249 options to -help for specific documentation of those options, like
252 $ scm -help do_not_drop
254 - Minor changes in how the [nm_version] section in psn.conf i
255 handled. Look in the psn.conf file that is distributed with PsN for
257 - SDE modelfiles are now supported. (see the -sde_records option)
258 - The update_inits method of the model class now accepts an output
260 - Some attributes and accessors of the PsN-Core class 'output' have
261 had their names changed to make it more clear what their purposes
263 finalgrad is now called final_gradient
264 parampath is now called parameter_path
265 gradientpath is now called gradient_path
266 covterm is now called covstep_termination
267 - The Utilties now saves the command line that started it in a file
268 "command.txt" under the corresponding directory.
269 - Added a "silent" option to the Utilties that quenches all output.
270 - "nice" option added to utilties that allows you to set priority in
272 - Rudimentary support for NorduGrid reinstated.
273 - Installer script rewritten. It now supports Perl installations in
274 nonstandard directories. Discrepancies between *nix and MSWin32 is
278 - The code that is autogenerated with "dia2code" is optimized a
279 bit and saves a hash "__valid_parameters" in each object, which
280 allows the constructors to be used as a generic (shallow) copy.
283 - continuous spelled correctly throughout SCM.
286 Changes from 2.1.5 to 2.1.6
289 - Stratify_on bug in bootstrap method fixed.
290 - Data files in directories different from model file are now
292 - Path name handling reimplemented. Relative paths are now handled
294 - Models that generate table files would make PsN complain about
296 - Its now possible to specify multiple output names as a
297 comma-separated list to execute.
298 - The scm can now go backwards.
299 - Handling of table files now responds to the
300 "ignore_missing_output_files" flag.
301 - nmtran error messages was lost in Windows 2K. Execution of nmtran
302 is more plattform independent now. It uses a temporary file, but
303 it is never bigger than the nmtran messages.
304 - Fixed a bug that limited choice of compiler under unix.
305 - Typo in scm utility fixed. (included_relations in backwards step).
308 - Added support for "$NONPARAMETRIC" record in model files.
309 - A few improved error messages related to parsing of model files.
310 - The SCM config file now features sections for parameter bounds and
314 - psn.conf settings are now available as global variables. psn.conf