Bug 40 - comments on Explanatory Supplement v54
comments on Explanatory Supplement v54
Status: NEW
Product: GALPROP Code
Classification: Unclassified
Component: Manual
54 (in WebRun)
PC Linux
: --- enhancement
Assigned To: Andy Strong
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2010-11-06 10:27 PDT by Andy Strong
Modified: 2010-11-10 14:20 PST (History)
4 users (show)

See Also:
Web Browser: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Andy Strong 2010-11-06 10:27:17 PDT
copied from Forum by user tomassetti:

Hello. I gave a look to the manual; the many new features of v54 are scattered throughout the document, so not very clear at first glance.
So, beside this new manual, I think it would be helpful a short summary document (just few pages) outlining all the differences v54/cv50 in few points, e.g.:
1) Physics processes. Is the description of CR physical processes (transport, interactions) changed? Where?
2) Physics inputs. Nuclear data files changes (which channel added), ISRF, HI, CO files. Quantities in E-loss routines.
3) Parametrizations. Differences in source spectra & spatial distribution, gas distribution, B-field models, cross section parametrization routines.
4) Technical aspects. (less important) Memory management or code organization; interfaces with eg. Galdef format.
Comment 1 Andy Strong 2010-11-06 10:32:51 PDT
thanks for the comments, this feedback on the Explanatory Supplement is very useful.
as before there is place for this under BUGS, goto galprop code and category Manual.
Meanwhile some answers...
I don't want to make a separate document since 'do everything in one place' seems a good motto.
But there can be special sections for what you suggest, e.g. summaries of differences.

You are right that the differences are scattered, just as they are throughout in code.
I thought they were most usefully described at the places in the document where they come up.
It would be hard now to list all the differences since v54 has been going on a few years now.
I will revise the relevant sections as far as possible, but one really should go now for v54 and not worry too much about everything that changed.

1) physics processes are not changed.

2) nuclear physics channels have been updated but not documented in the Exp.Supp. There is a paper which can be referenced meanwhile.
the ISRF changed at lot, but has not been written up yet.
HI, CO changed, not yet written up.
E-loss routines are unchanged.

3) some of these are documented, some not.

4) a start has been made on the code architecture, at a high level.
also some more very detailed description of the more important propagation routines has been added.
for the GALDEF parameters, this is done per parameter which I think is most useful.

Note that documentation is a never-ending task especially with code being steadily updated, and people find it more fun to write code than to write it up.


NB many of the changes affecting the user are documented in real time in the README file shipped along with the code. the nitty-gritty changes are seen by the developers via svn.
Comment 2 Andy Strong 2010-11-06 10:35:06 PDT
originated in Forum here:
http://galprop.stanford.edu/forum/viewtopic.php?f=19&t=113&p=303#p303
Comment 3 Andrey Vladimirov 2010-11-06 11:01:57 PDT
There is also a brief list of changes in this paper: http://arxiv.org/abs/1008.3642

Shared-memory parallel support with OpenMP to take ad- vantage of multi-processor machines;
• Memory usage optimization;
• Implementation of the HEALPix output of gamma-ray and synchrotron skymaps. The HEALPix format [38] is a stan- dard for radioastronomy applications, as well as for such instruments as WMAP, Planck etc.;
• Implementation of the MapCube output for compatibility with Fermi-LAT Science Tools software [39];
• Implementation of gamma-ray skymaps output in Galac- tocentric rings to facilitate spatial analysis of the Galactic diffuse gamma-ray emission;
• More accurate line-of-sight integration for computing dif- fuse emission skymaps;
• 3D modeling of the Galactic magnetic field, both regular and random components, with a range of models from the literature, extensible to any new model as required;
• Calculations of synchrotron skymaps on a frequency grid, using both regular and random magnetic fields;
• Improved gas maps, which are computed using recent H I and CO (H2 tracer) surveys [40, 41], with more precise assignment to Galactocentric rings;
• A new calculation of the Galactic interstellar radiation field using the FRaNKIE code (Fast Radiation transport Numerical Kode for Interstellar Emission, as described in [17]) and implementation of the corresponding changes in GALPROP;
• Considerably increased efficiency of anisotropic inverse Compton scattering calculations;
• GALPROP code is compiled to a library for easy linking with other codes (e.g., DarkSUSY [42], SuperBayeS [43]);
• Numerous bug fixes and code-style improvements;
• Improved configuration management via the GNU auto- tools. Multiple *NIX system and compiler targets (gcc, intel, llvm, open64) are supported;
• Bugzilla available at the GALPROP WebRun URL for user-submitted bug tracking.
Comment 4 Andy Strong 2010-11-06 11:07:41 PDT
thanks for reminding me, will put those in the document and also reference to that paper.
Comment 5 Nicola Tomassetti 2010-11-06 11:09:28 PDT
Hi. Please don't take my post as a request. I just wanted to give you my impressions in reading the manual. 
Since v54 is not public (the source code) but publicly runnable by users (as a black-box), I think documentation becomes now more important than for the previous versions (in which many answers were given by the code itself). 
An analytical v54/v50 differences list would supply this lack.
Comment 6 Andy Strong 2010-11-06 11:17:38 PDT
we welcome suggestions/comments and even requests!

the code IS available on request (see under CODE) and even if you run WebRun it is useful for reference precisely to avoid the 'black-box' effect.

But the documentation is something we have to put more effort into as well of course.
Comment 7 Andy Strong 2010-11-10 00:24:34 PST
re: http://arxiv.org/abs/1008.3642
is this paper now accepted so I can put the full reference ?
Comment 8 Andy Strong 2010-11-10 05:05:03 PST
added reference to http://arxiv.org/abs/1008.3642
and copied the enhancements list from there to Section 8.
svn r594
Comment 9 Andrey Vladimirov 2010-11-10 10:03:17 PST
As of today, the last status update is dated October 14 and says "Under Review" (had been "Handled by Editor" before that).


(In reply to comment #7)
> re: http://arxiv.org/abs/1008.3642
> is this paper now accepted so I can put the full reference ?
Comment 10 Andy Strong 2010-11-10 10:28:56 PST
re paper: amazingly long, submitted in August !
Comment 11 Andrey Vladimirov 2010-11-10 14:20:47 PST
Well, submissions to this journal are free, and, as long as there is some progress, I wouldn't complain that they are slow. I will wait another week and send them a message.

(In reply to comment #10)
> re paper: amazingly long, submitted in August !