[POTKU] New version for Potku, multiple bugs corrected, please start using the newest version

Laitinen, Mikko mikko.i.laitinen at jyu.fi
Wed Apr 15 14:56:59 EEST 2020 

Dear users of Potku software,

In the Potku www-pages:
https://www.jyu.fi/science/en/physics/research/infrastructures/accelerator-laboratory/pelletron/potku/potku-analysis-software-for-tof-erda
there's now an update of the recent progress.

The new compiled Windows and MacOS versions can be downloaded from here:
https://www.jyu.fi/science/en/physics/research/infrastructures/accelerator-laboratory/pelletron/potku/release_versions

Although the same text and info can be found from the pages above, I'll give the progress info also directly on this email below:
------
2020 April (10.4.2020) update! From the beginning of 2020 we have had a IT student Juhani Sundell hunting for the multiple bugs and making more permanent fixes to the code. Now the isotope efficiency related bug(s) is corrected and the MCERD side seems to work quite fluently.  Also, the simulated energy spectra are now updating instantly when one changes the recoil distribution and the foil-related mismatch between slab-analysis and MCERD side is one of the corrected issues.

However, user can still do wrong if for example the efficiency file is not named correctly, meaning file name has spaces or special characters (H1.eff, H2.eff are default names but H1_Jan_2017.eff -type format works also), or for example user tries to do depth profiles from "H" + "1H" + "2H" isotopes _at the same time_. Here, result would be that H and H1 concentration would be summed up as to 1H, with correct efficiency's, and 2H would be ok with correct efficiency -> this is due to user error, but we'll work to prevent these also. Automated fitting tool in MCERD-side is still to be developed more user friendly and still has some bugs.

ALSO: After the summer 2020, one can expect new ePotku, the www-browser/server based version of the desktop Potku, where your analysis stay on the the secure server, and you can continue your analysis from anywhere, including continue your offline analysis, after uploading the date to your own secure folder. One can also share the data and analysis to your own research group, and beyond if needed. In addition to normal private sharing ePotku will in the future provide Open Access platform to your Open Data. So this update is also currently developed, on the parallel track with the desktop version.


  *   In Windows version, just start from potku.exe after unpacking.
  *   Mac version need to be started from the command line, not from the icon.
Tcl/Tk-software is needed on Mac (should be installed as default). Mac version runs through the terminal by running a command "./potku" in potku folder.
  *   Linux and other users can found the Potku from the github, but take the "external files" from the win/mac packages. (we'll update the linux compiled version soon)


---- Some of the corrected bugs - and need to know 'bugs" or "features":

  *   Efficiency files for multiple isotopes should now work from now on (the issue earlier was related to so called "external files" which were not updated at the summer 2019 version vs December 2018).
  *   User can still do wrong if for example the efficiency file is not named correctly (has spaces or special characters -> H1.eff, H2.eff as default but H1_Jan_2017.eff -type format works also), or for example user tries to do depth profiles from "H" , "1H" and "2H" isotopes _at the same time_. Here result would be that H and H1 concentration would be summed up as to 1H, with correct efficiency's, and 2H would be ok with correct efficiency -> this is due to user error, but we'll work to prevent these also.
  *   Earlier, if user changed the efficiency files multiple times during the analysis, the new additions did not register properly without restart. Now they do work and user can change, test, multiple efficiency corrections fast in a single analysis.
  *   If one used zoom in the depth profiles to draw selection limits (blue vertical lines), the window unzoomed. Now this is fixed for depth profiles at least, will be fixed elsewhere also.
  *   Might be still an issue: Long filenames/long chain of folders might cause troubles to 'gsto' and 'stopping' C-codes still. (possible if total folder path of the working files is longer than 256 or so characters..)
  *   "Open Recent requests" - option is now in the file-tab, which can be used when starting potku and looking for previously opened requests.
  *   MCERD: there were a critical mismatch of the first timing foil handling when using the MCERD and compared the data with slab analysis (measured) energy. For the energy spectra, the slab-analysis side _adds_ by default the energy loss to the energy which is actually detected between timing gates(ToF) -> this output (E+E_loss_in_foil) then matches to the original energy from the sample. However, before the fix, MCERD used this slab-analysis output energy for it's own comparisons, resulting that front edge of scattered ion energy with thick T1 foil resulted: slab analysis got very high original energy from the sample, MCERD (as detector energy) resulted very low energy. Now the slab analysis energy, the measured-one, is brought to the MCERD-side without the energy loss-addition in the T1. Now both programs compare the ToF(energy) that is seen (eg. recorded vs simulated) by the difference of timing signals of two timing gates.
  *   MCERD: when one adds processing units, the MCERD now divides the original "number of events" to be simulated with this number, to each of the core. Eg. simulated ions + presimulation = 1 000 000 + 100 000  and if one uses 4 cores, each core currently calculates 250 000 + 25 000. (previously this was 4x 1 000 000 + 4x 100 00, which wasn't speeding up anything.) HOWEVER: presimulation minimum number should not be much smaller than 10 000 at least, this is now by the user to be seen (in the console-window) but we'll work automation for this too.
  *   MCERD: one can now select some of the elements "OFF" from the "calculate percentages", to reject uncertainty tests or so ending to the total concentration.
  *   MCERD: No new graph is needed if the recoil distribution is changed, the energy graph updates in about 0.3 seconds after the spheres/dots are changed in the recoil distribution.
  *   MCERD: there's a bug that crashes the potku, in the recoil[Image preview] atom distribution, if opening the menu from where element's name or color can be changed, anf in one presses "OK" potku will crash (see pic). (this is already fixed at the github source code).
  *   Optimization results can now be transferred to recoil distribution, but this is work in progress as, one needs to restart that those _really_ appear there. Also the optimization starts the MCERD still even if there's correct amount of MCERD events.. will be fixed too..
  *   In the sample data-folder<https://www.jyu.fi/science/en/physics/research/infrastructures/accelerator-laboratory/pelletron/potku/sample_data>, example_efficiency_files.zip, the 2H.eff file "X-column" had turned to date-format for energies above 1 MeV. This is now corrected for the sample data.  NOTE it needs to be checked that windows tab-delimited MeV, efficiency - columns work on the mac, where space delimited columns are often used.. this applies to unix users also. Tab and space delimited versions are provided for test.


So now it's a very good time to test the new versions - and to give feedback and report findings so that they could be possible adapted to the new ePotku addition also!

Please, share your thoughts and ideas with us - and stay safe!

Best regards,

      Mikko.



--
Mikko Laitinen, Researcher
University of Jyväskylä
Department of Physics
________________________________________________________
Postal address: University of Jyväskylä,
Department of Physics, P.O. Box 35,
40014, Finland
Street address: Survontie 9 D, Ylistönrinne, Jyväskylä

Office: +358 40 8054 113
Personal: +358 400 99 48 36
Telefax: +358 14 617 411
E-mail address: mikko.i.laitinen at jyu.fi

What was necessary, was done by yesterday;
We're currently working on the impossible;
For miracles, we ask for a 24 hours notice ... :)

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.jyu.fi/pipermail/potku/attachments/20200415/16f7f8bd/attachment.html>

More information about the Potku mailing list