| blob | 2fc406310baf190f299ef9fec99aca3dedfed321 |
1 This version of PsN is tested with perl-5.8.8
2 Depending on your distribution you may have to install the following
3 perl packages:
5 Math::Random
6 Storable::Store
8 All available from CPAN ( www.cpan.org )
10 Changelog
11 ----------------------------------------------
13 Changes from 2.2.4 to 2.2.5
15 Features
17 - mcs script renamed to sse
18 - sse calculates lots of nice results.
20 - msfi/msfo.
22 If the model allready has the MSFO= option or the $MSFI record set
23 and one of the files exists, then that file will be used as $MSFI
24 file in the first run. If both files exists, then the MSFO file
25 will be used as MSFI file. The file will be renamed psn_msfo-0.
27 The MSFO option will be set to psn_msfo.
31 Bugs
33 - Multiline ETABAR and PVAL parsed correctly.
35 - Modelfit did not copy extra subroutine files in some cases.
37 - <new script> Add IGNORE=@ to estimation models unless it
38 exists. Necessary for ignoring headers in table output.
40 - Memory leak in output parser plugged. Output files from simulations
41 could use vast amounts of memory.
43 - Bootstrap options "skip_minimization_terminated" can now be disabled
44 by typing "-no-skip-minimization_terminated" (or -noskip-...)
46 - cdd xv can no be disabled with -no-xv or -noxv
48 - mc tool no uses considerably less memory. It is a lot slower
49 because it reads a lot of unecessary stuff from disk, it should be
50 reviewed.
52 - Memory leak in modelfit due to intermediate "raw_results.csv" in
53 NM_runX directory fixed by removing the feature, should be removed.
55 - A initial estimate of form: $OMEGA BLOCK SAME would not be
56 recognized by PsN model file parser. This is now handled
57 correctly. In 2.2.X it can be avoided by typing BLOCK(X) SAME,
58 where X is the size of the block.
60 Changes from 2.2.3 to 2.2.4
62 Features
64 - "--compute_cwres" option renamed to simply "--cwres"
66 - Simulated data suitable for creating mirror plots can be generated
67 by using a new options called --mirror_plots. The option takes a
68 value which defines the number of simulations that will be
69 geneated.
71 - A new option, --iofv, enables the computation of individual
72 objective function values. These values are printed to a table file
73 called iotab*, where * is the same number as is defined for any
74 sdtab or patab tables. These names correspond to the normal Xpose
75 table file format. If no sdtab or patab table is found in the
76 NONMEM control stream, the iotab file is given number 1.
78 - msfo resumes. Basic support for automatically rerunning a model
79 that has the MSFO=msffile option on $ESTIMATION. If the option is
80 set, the "msffile" exists and the option --msfo_resume is given,
81 PsN will remove initial estimates, and add the $MSFI record. This
82 is very untested and if you have OMEGA blocks, the model will run
83 but PsN will be unable to read the output file properly.
85 - "no_remote_compile" and "no_remote_execution" options has been
86 removed, as they were quite seldomly used and added a lot of
87 complexity to parallization code.
89 - Added --prepend_model_file_name option which adds the model file
90 name to output files to prevent them from being overwritten.
92 - A new option called "crash_restarts" which controls the number of
93 times PsN will restart a crashed run. (Simply put, run is
94 classified as crashed if the output file looks like it is cut off)
96 - PsN will now automatically run R-scripts generated for the
97 bootstrap, cdd and llp if R is installed and configured.
99 - When the CONT data item is used - either defined by the user or
100 automatically through the -wrap_data option - the tables get a
101 weird format where the real table content is printed on CONT=0 rows
102 and nonsense data printed on CONT=1 rows. A new option called
103 -unwrap_table_files reformats the table files and removes the CONT
104 column as well as all CONT=1 rows.
106 - update_inits script that will put final estimates from a NONMEM
107 output file and put into a given model file. This is quite
108 experimental. PsN will reformat the model file quite lot and
109 comments in the file might get lost. Your model file will be copied
110 with the addition of a ".org" extension, so you wont loose your
111 model.
113 Bug Fixes
115 - SGE queue and resource options were ignored. Fixed thanks to Jeroen
116 Elassaiss and Stefan Verhoeven.
118 - LSF Monitor fixed to work with the new serialized structure of PsN,
119 again thanks to Jeroen and Stefan.
121 - CWRES table file number now defaults to 1 if there is no patab or
122 sdtab to take the number from.
124 - Fixed run bugs under Windows 2000.
126 - Better support for G77 under windows.
128 - Parser for output files improved
130 Changes from 2.2.2 to 2.2.3
132 Features
134 - There has been a lot of requests for a raw_results file for the
135 runs in the NM_run catalogs, i.e. a summary of the runs with
136 perturbed initial estimates. This is now implemented. One
137 raw_results file is now created in each NM_run* catalog.
139 - The R-scripts for graphical visualization of bootstrap and cdds
140 have been updated. The format for the graphs is now pdf instead of
141 postscript. The bootstrap script can now be adjusted to exclude
142 bootstrap samples that contain copies of a given set of
143 individuals.
145 - When the wrap_data option is used, the secondary columns, i.e. the
146 data columns that are kept on rows with CONT=1, are now available
147 in $ERROR.
149 Bug Fixes
151 - The calculation of CWRES was previously done on the _last_ run of a
152 set of retries, not on the _best_ run. This has now been corrected.
154 - The restarting of crashed runs was broken. This is now fixed.
156 - The asynchronuous job scheduling broke the handling of seed numbers
157 for random sequences (e.g. perturbation of initial estimates). This
158 is now fixed. The present version is however not compatible with
159 older versions and old runs should therefore not be resumed using
160 this version.
162 - wrap_data should now work as it should. The functionality was
163 broken as a consequence of the move to the asynchronuous job
164 handling.
166 - Labels for omegas were not handled correctly for SAME blocks. Among
167 other things, this had the effect that the updating of initial
168 estimates from the original model to the method created models
169 (e.g. bootstrap samples) was corrupted. This is now fixed.
171 - The transformation of the estimate of the omega matrix from
172 variance and covariances to standard deviations and correlations
173 was rewritten to address problems with a combination of BLOCK(X)
174 ans SAME where X was larger than 1.
176 - The selection of a 'best' model fit after a series of perturbations
177 of the initial estimates will be changed. In 2.2.2, the lowest OFV
178 among fits of the perturbed models was collected across all models
179 (still only comparing identical models with perturbed initial
180 estimates). In 2.2.3 the lowest OFV is collected in three
181 categories of model fits: those that minimized succesfully, those
182 that did not minimize successfully but returned an OFV plus an
183 estimate of the number of significant digits and finally failed
184 runs that managed to return an OFV but not much more. This will
185 give PsN a more fine-grained control over the results. For example,
186 it would be desirable to be able to return the run with the lowest
187 OFV within the selection of runs with a successful minimization
188 even though the lowest OFV of all perturbed runs was returned from
189 a run that terminated.
191 - The sigmas were not parsed correctly when tables were printed to
192 the output file. This is not fixed.
194 - Terminated covariance steps were sometimes still being parsed (and
195 failed). Fixed.
197 - Setting the clean level to 3 did not remove the
198 NM_run-directories. This is now fixed.
200 - The parsing of options for the initial values of omegas and sigmas
201 sometimes wrongly recognized letters behind a semi-colon (i.e. a
202 comment) as keywords for a block (FIX, SAME, etc). This is now
203 corrected.
205 - If you "fix" more than one omega or sigma in a block PsN would put
206 a FIX after each value, which NONMEM doesn't allow. Now PsN will
207 only put one FIX inside a block. However, if a any part of the
208 block has been fixed, its likely so that PsN can not(!) unfix
209 it.(so, one bug fixed, another introduced)
211 - If you an initial estimate was added with "_init_attr", the option
212 "add_if_absent" was true and the last initial estimate was "SAME"
213 PsN would try to add the estimate after "SAME" which is wrong, now
214 PsN adds a new record.
216 - Fixed a parsing error of the lst file. If simulation without a
217 SIGMA was performed, the parser failed to read OMEGA initial
218 estiamtes.
220 - The internal method "indexes" would not properly handle "SAME" in
221 initial estimates. The result was that "update_inits" would skip
222 initial values, and get them out of order. Fixed.
225 Changes from 2.2.2-beta4 to 2.2.2
227 Features
229 - CWRES can now be calculated automatically if you have R and Xpose
230 installed. Remember to add a definition of IPRED in your NONMEM model.
232 - This version handles non-PsN standard (or absent) suffixes for
233 model files better than earlier versions.
235 Bug Fixes
237 - MSFO files are now copied back to the same directory of the model
238 file.
240 - MSFI files are now automatically copied to the PsN NONMEM run
241 directories.
243 - Output files from runs where the $MSFI record was used were not
244 parsed correctly. This is now fixed.
246 - Output files from runs where no $OMEGA _or_ $SIGMA were defined
247 were not parsed correctly.
249 - The header and the results from problematic runs, i.e. failed
250 covariance step, warnings etc, were not handled correctly when the
251 raw_results files were compiled. This is now fixed.
253 - In 2.2.1, we added a check for a '1' to find the end of the omega
254 and sigma estimate areas. this was a mistake since problems with
255 more than 13 omegas or sigmas get a '1'-line inserted before the
256 14:th estimate (for some reason). This is now removed.
258 Changes from 2.2.2-beta3 to 2.2.2-beta4
260 Features
262 - No new features
264 Bug Fixes
266 - Adaptive number of threads is now working in the new asynchronuous
267 job handling
269 - The CWRES computation had some fixes
271 - $INFN is now printed before $MODEL
273 - $OMEGA is now printed befor $PK
275 - PRIOR is now handled in the output module. It is still a bit
276 problematic to parse PRIOR output files correctly. Use with
277 causion.
279 - The initial estimates of large omega and sigma matrices is now
280 better handled
282 - The order of parsing events in the output module is now working
283 better.
285 - The handling of retries (or the results of retries) is now working
286 as it should.
288 Changes from 2.2.2-beta2 to 2.2.2-beta3
290 Features
292 - CWRES: Conditional weighted residuals. The necessary verbatim code
293 for the computation of CWRES are now automatically added if the
294 option -compute_cwres is used.
296 - PRIOR: The use of priors in NONMEM VI is now supported in the
297 computer intensive methods
299 - The behavior of the parallel execution has been revised, moving
300 from a synchronuous to an asynchronuous solution. Previously, the
301 memory foot-print of a parallel execution of say 100 NONMEM jobs
302 was huge (counted in several GB) and did not scale well. Now, it
303 will very seldom exceed 100 MB and more or less keep that size,
304 regardless of the number of parallel jobs.
306 Bug Fixes
308 - PsN couldn't parse IGNORE=(ITEM=VALUE) options in $DATA. This is
309 (again) fixed. It was previously corrected in the development
310 branch of PsN but never transferred to the stable branch.
312 - Fixed SAME block correlation computation.
314 Changes from 2.2.2-beta1 to 2.2.2-beta2
316 Features
318 - The parsing of the NONMEM output files has been revised and a
319 strict check is now performed after each section has been parsed. A
320 parsing error is raised if any non-compliant structures are
321 found. By doing this we get a better idea of which runs that really
322 finishes and which runs that are pre-maturely terminated
323 (e.g. through a operating system error). Remember that this is a
324 beta release and that there could still be valid text structures
325 that are not yet recognised correctly by PsN.
327 - Support for United Devices grip MP has been added and is available using
328 "-run_on_ud" command line option.
330 Bug Fixes
332 - the options directory and backwards_dir can now be specified in the
333 scm configuration file
335 - It is now possible to skip the definition of either continuous and
336 categorical covariates in the scm configuration file
338 - The exponential relation is now added correctly in the
339 scm. Previously it was incorrectly added as (for weight on volume)
340 TVV = THETA(1)*(1+EXP(THETA(2)*(WGT-median(WGT))))
341 Now it is added as
342 TVV = THETA(1)*EXP(THETA(2)*(WGT-median(WGT)))
344 - The output parse now computes correlations for omegas and sigmas correctly
345 when you have SAME blocks.
347 Changes from 2.2.1 to 2.2.2-beta1
349 Features
351 - The number of significant digits per parameter after a terminated
352 minimization are now parsed.
354 Bug Fixes
356 - The sample size option of the bootstrap did not work correctly with
357 stratified resampling. This is now fixed.
360 Changes from 2.2.0 to 2.2.1
362 Features
364 - Added support for the Intel Fortran compiler (mostly for those
365 using the new Intel Macs).
367 - We have added two new model fit quality terms: 'terminated' and
368 'really_bad' to the previous 'normal' and 'picky'.
370 picky - Minimization and covariance steps successful, no
371 parameters near boundaries and no warnings from the
372 covariance step.
374 normal - Minimization successful
376 terminated - Estimates of the objective function value and the
377 number of significant digits are available.
379 really_bad - An estimate of the OFV is available (but usually not
380 much more...).
382 Given that more than one run from a run has been performed (with
383 different initial estimates but otherwise identical) the following
384 selection procedure is done:
386 The lowest OFV of all tries is collected. Then, all runs within
387 5 units of the lowest OFV is tried according to
389 1. Of the runs passing the picky criteria, choose the run with the
390 highest number of significant digits.
392 2. If no run has passed the criteria above: of the runs passing the
393 normal criteria, choose the one with the highest number of
394 sginificant digits.
396 3. If no run has passed the criteria above: of the runs passing the
397 terminated criteria, choose the one with the highest number of
398 sginificant digits.
400 4. If still no run has passed the criteria above: of all runs that
401 at least have an estimate of the OFV, choose the run with the
402 lowest OFV.
404 5. As a last resort, choose the first run.
406 Comments: This is of course a very pragmatic approach but it is
407 needed to make the automatic procedures of the PsN tools to
408 work. We cannot force the user to manually scrutinise the results
409 of every single run of for example a bootstrap. The user has the
410 opportunity to do so if he or she wishes. If you do not feel
411 comfortable with this selection procedure or if you have comments
412 on how to improve it, please contact me (Lars Lindbom) at
413 lars.lindbom (at) farmbio.uu.se.
415 Bugfixes
417 - Data files with spaces around commas used for columns separation
418 could be problematic to parse. This is now fixed.
420 - There was a serious error in the seed handling in version
421 2.2.0. Under the following circumstances, the seed number was
422 ignored and a default seed number used: If the output file from the
423 original run was not present in the same directory as the original
424 model file when a PsN tool was started. This affected all PsN
425 tools, except the execute tool. This is now fixed.
427 - Up to 20 omegas can now be parsed in the output file. This is still
428 a limitation and it will be addressed in later releases.
430 - On fast distributed systems with a shared file system, the massive
431 parallelisation of the PsN tools could cause synchronization
432 problems (files could sometimes not be found, read or written). The
433 root of this problem is somewhat outside of the scope of PsN but we
434 have tried to find a workaround for this for the LSF grid system.
436 Changes from 2.1.10 to 2.2.0 ( released for PsN course )
438 Features
440 - PsN Output module use a proper Matrix library for handling some of
441 the matrices of the outputfile.
443 - The bootstrap can now be resumed with a lower number of
444 samples. Meaning that if you are in a hurry and want to stop the
445 bootstrap and get results after 1000 out of 2000 runs, you can. At
446 a later time you can then resume the 2000 runs.
448 - Nonmem warnings are now printed if "--verbose" is givven.
450 - Automatic incrementation of directory names are a bit smarter (no
451 more "directory modelfit_dirX is in the way" messages).
453 - Clean is much more stringent and intelligent and now manages to
454 remove entire tool directories if desired.
456 - option list printed from "-h|-?" only shows utility specific options.
458 - Default values configurable on a per utility basis.
460 - Online documentation improved
462 - Parallel run of NONMEM is not aborted as before if a outputfile is
463 not correctly parsed. It is just ignored, as if it had not
464 minimized successfully. (actually it is the output parser that
465 doesn't die, but the effect is the same)
467 - Parallel run is a bit nicer to the system when compiling many
468 NONMEMS. By adding a configurable delay between starts.
470 - Matlab and R scripts for various plots from cdd and bootstrap
471 results are distributed with PsN.
473 - Vast improvments in results presentations from all tools.
475 - Improvements in rerun algoritm with new "min_retries" option.
477 - Output from running utilties are much nicer.
479 - new "-quick_summary" prints minimization message and ofv value for
480 each finished run. As does "-verbose" option which also prints
481 "R:X" for each retrie (X is the retrie number).
483 - new "-summary" options prints diagnostics about the tool run when
484 it is finished. For "execute" the "-summary" option will have the
485 same function as the "sumo" utiltiy as well as some diagnostics.
487 - PsN now recognizes that control files may require fortran files and
488 copies them to the run directory. No need to add fotran files to
489 the --extra_files option any more.
491 - Added --extra_output which allows you to copy more than the output
492 file from the run directory.
494 Bugs
496 - PsN couldn't parse IGNORE=(ITEM=VALUE) options in $DATA. Fixed :)
498 - If star notation was used in the scm config file section [inits],
499 [upper_bounds] or [lower_bounds] there was a risk of defining
500 relations that was not specified in [test_relations]. Which in turn
501 would cause a lot of diffrent errors. Fixed :)
503 - Using low_cutoff and hi_cutoff in data::sd resulted in PsN trying to use
504 the new compact data strings as the old arrays. fixed.
506 - Removed unecessary regular expression in output file parser.
508 - Not really a bug. But i tried to add "FIX" as a record option to a theta
509 records. Then when "store_init" was called on the "FIX" option PsN crashed
510 because the option was not a "theta_option". So i added a check for
511 "store_init" abilties... if FIX would become a record option in the future.
513 - Member accessor "datafile" replaced with "datafiles" in model. Affected
514 only the bootstrap which have been updated.
516 - An assumptions about outputfiles crashed sumo occationally. Some more
517 rigorous contentchecking has been added and fixed the problem.
519 Internal changes
520 - Cleanup of dubious utility options.
521 - Option help sorted alphabetically.
522 - "flush", "target" removed from model.
523 - "remove_temp_files" removed from modelfit.
524 - Changed the Makefile considerably, much simpler now.
525 - added "column_to_array" function in data module.
527 Changes from 2.1.9 to 2.1.10
529 Features
530 - The 'picky' option of all utilities now triggers on
531 COVARIANCE STEP ABORTED and
532 PROGRAM TERMINATED BY OBJ
534 Changes from 2.1.8 to 2.1.9
536 Features
537 - Nordugrid support reintroduced (but will probably not be developed further)
538 - Sumo now supports multiple problems and subproblems.
539 - Sumo has new option "precision" which control the number of digits
540 to present estimates with.
541 - An extra 'short' logfile is output from the scm with the most important
542 information.
544 Bugs
545 - Bootstrap option "stratify_on" doesn't work with columns defined in
546 extra data files given with "extra_data" option. Not Fixed.
547 - Resumes of scm's with extra_data_files is now possible. Still some
548 problems with scm's where search_direction=both.
549 - LSF options was not communicated properly between tools and subtools
550 so submission of bootstraps, llp, cdd and scm did not work. Fixed.
551 - Default values for LSF can now be configured in psn.conf.
552 - Print order of modelfiles places $MIX before $PRED. Fixed.
553 - extra_files option was broken. Fixed.
555 Internal Changes
556 - Tool will always fork subtools, this will save memory in the scm (and
557 other recursive tools).
558 - Common options need only be specified once. In one of three categories:
559 model_options, tool_options, script_options.
560 - models can now be created with "reference" objects. This solved the
561 scm resume problem.
562 - Added function to model that cleans extra data code.
564 Changes from 2.1.7 to 2.1.8
566 Features
567 - Rudimentary support for execution on Platform LSF grid systems is now
568 included. The following extra attributes are available to all tools
569 for controlling the execution:
571 run_on_lsf - Use LSF
572 lsf_queue - Use this queue on the grid
573 lsf_options - General options for the grid queuing system.
574 lsf_project_name - Optional. May be used by the grid system
575 administrator to register control project
576 cpu usage.
577 lsf_job_name - Optional. May be used by a grid system
578 administrator to control multiple runs on a
579 grid system.
581 - All tools now have a raw_results file holding the diagnotics and
582 parameter estimates for each NONMEM run.
583 - All tools now handles model files containing one $PROBLEM and no
584 SUBPROBLEMS only.
585 - The Bootstrap and CDD routines now create separate files holding
586 information about which individuals that were included/excluded
587 in/from each created data set.
588 - The Bootstrap results file now has warnings section to better
589 display which runs that were used for result calculations and which
590 were not
591 - PsN now tries to read psn.conf from the users home directory. (Will
592 require some extra documentation).
593 - Optimized modelfit thread operation by removing Storable::Store.
594 - The clean options has four levels of operation.
596 0 - No cleaning at all.
597 1 - Clean more, but make sure resumes work (this is default).
598 2 - Clean more, but make sure summarize work.
599 3 - Clean everything, keeping only results and logs.
601 - Some attributes and accessors of the PsN-Core class 'output' have
602 had their names changed to make it more clear what their purposes
603 are:
604 termination is now called minimization_successful
605 termination_message is now called minimization_message
606 covstep_termination is now called covariance_step_successful
608 these attributes are new:
610 covariance_step_warnings (boolean; 0 or 1)
611 estimate_near_boundary (boolean; 0 or 1)
613 The attribute which was previously called termination could have the
614 values 0, 0.5 and 1 for 'minimization terminates', 'minimization
615 successful but warnings or errors were printed from the covariance
616 step' and finally 'minimization successful'. The new attribute
617 'minimization_successful' can only take the values 0 and 1 and the
618 new attribute covariance_step_warnings now takes the value 1 when
619 covariance step warnings were found. Checks are also perfomed to
620 test whether the estimates are close to any boundary.
621 - Html documents is now generated from command line help text.
622 which ensures that they are consistent and up to date.
623 - Personal configuration file now available under Unix like OS:s
625 Bugs
626 - Options specific to modelfiles were ignored in some Utilties.
627 - Gradients now output in windows. Just like with nmfe.
628 - Bootstrap now consumes considerably less memory.
629 - Removed unused/obsolete "resume" function from modelfit.
630 - Sorted method entries in modelfit diagram.
631 - Fixed bugs in parsing of OFV and eigenvalues.
633 Changes from 2.1.6 to 2.1.7
635 Bugs
636 - "clean" and "compress" options now work under Windows.
637 - SCM now handles crashed runs appropriately.
638 - The ofv_forward, ofv_backward, p_forward and p_backward options of
639 scm now works as they should.
641 Features
642 - SCM improvements. The final model is now available in the
643 first-level scm directory of each step.
644 - The Utilties -help option is improved. You can now give a list of
645 options to -help for specific documentation of those options, like
646 so:
648 $ scm -help do_not_drop
650 - Minor changes in how the [nm_version] section in psn.conf i
651 handled. Look in the psn.conf file that is distributed with PsN for
652 more information.
653 - SDE modelfiles are now supported. (see the -sde_records option)
654 - The update_inits method of the model class now accepts an output
655 file name as input.
656 - Some attributes and accessors of the PsN-Core class 'output' have
657 had their names changed to make it more clear what their purposes
658 are:
659 finalgrad is now called final_gradient
660 parampath is now called parameter_path
661 gradientpath is now called gradient_path
662 covterm is now called covstep_termination
663 - The Utilties now saves the command line that started it in a file
664 "command.txt" under the corresponding directory.
665 - Added a "silent" option to the Utilties that quenches all output.
666 - "nice" option added to utilties that allows you to set priority in
667 unix.
668 - Rudimentary support for NorduGrid reinstated.
669 - Installer script rewritten. It now supports Perl installations in
670 nonstandard directories. Discrepancies between *nix and MSWin32 is
671 smaller.
673 Internal changes
674 - The code that is autogenerated with "dia2code" is optimized a
675 bit and saves a hash "__valid_parameters" in each object, which
676 allows the constructors to be used as a generic (shallow) copy.
678 Misc
679 - continuous spelled correctly throughout SCM.
682 Changes from 2.1.5 to 2.1.6
684 Bugs
685 - Stratify_on bug in bootstrap method fixed.
686 - Data files in directories different from model file are now
687 possible.
688 - Path name handling reimplemented. Relative paths are now handled
689 better.
690 - Models that generate table files would make PsN complain about
691 missing files.
692 - Its now possible to specify multiple output names as a
693 comma-separated list to execute.
694 - The scm can now go backwards.
695 - Handling of table files now responds to the
696 "ignore_missing_output_files" flag.
697 - nmtran error messages was lost in Windows 2K. Execution of nmtran
698 is more plattform independent now. It uses a temporary file, but
699 it is never bigger than the nmtran messages.
700 - Fixed a bug that limited choice of compiler under unix.
701 - Typo in scm utility fixed. (included_relations in backwards step).
703 Features
704 - Added support for "$NONPARAMETRIC" record in model files.
705 - A few improved error messages related to parsing of model files.
706 - The SCM config file now features sections for parameter bounds and
707 user-defined code
709 Internal changes
710 - psn.conf settings are now available as global variables. psn.conf
711 is only parsed once.