GALPROP
https://galprop.stanford.edu/forum/

Difference results from webrun and locally compiled code
https://galprop.stanford.edu/forum/viewtopic.php?f=7&t=197
Page 1 of 1

Author:  ranlu [ Thu Jan 29, 2015 2:38 pm ]
Post subject:  Difference results from webrun and locally compiled code

I am trying to reproduce the results in 1202.4039. I tried the webrun and the match between the two are very well. However the code I compiled on my machine generate very different results. The model I am testing is SNR_z4kpc_R20kpc_Ts150K_EBV5mag, I downloaded the galdef file from galprop website. Because I cannot find rbands_co10mm_v2_2001_qdeg.fits.gz, I use rbands_co10mm_v3_2001_hdeg.fits.gz instead, this is the only change I made. You can see from the log files I uploaded there is no other obvious difference between the webrun and my own run. But pi0 skymap is very different. In my own run there seems to be very few structure/contribution from the high latitude. Did I miss something obvious? I cannot do further testing since I do not have the v2 COR file.

Attachments:
File comment: local and webrun skymap at 185 MeV
pi0.pdf [856.96 KiB]
Downloaded 529 times
File comment: localrun log
galprop_log3.txt [234.11 KiB]
Downloaded 630 times
File comment: webrun log
log_54_webrun.txt [234.02 KiB]
Downloaded 544 times

Author:  gudlaugu [ Thu Jan 29, 2015 3:32 pm ]
Post subject:  Re: Difference results from webrun and locally compiled code

The problem is that the two gas maps you are using are not compatible. The co map is split into 9 annuli while the hi is split into 17. There is unfortunately no check for this inconsistency. If you really need to reproduce the numbers of the papers locally you could merge some of the annuli in the HI map to make them consistent. Otherwise running with the HI map provided with the code is advised.

Author:  ranlu [ Thu Jan 29, 2015 9:28 pm ]
Post subject:  Re: Difference results from webrun and locally compiled code

Thanks a lot for the explanation, I tested with rbands_hi12_v5_hdeg_zmax1_Ts125.fits.gz and now the output of galprop looks a lot better. I do not know the differences between this and the HI files in the addon package, since the galprop-54_data_111612_addon.tar.gz is released much later I assumed they are more "correct"/better. Is there any plan to release a qdeg CO map to be used with them, or how can I get one? Otherwise the HI maps from the addon package are not really useful.

Author:  gudlaugu [ Fri Jan 30, 2015 5:01 am ]
Post subject:  Re: Difference results from webrun and locally compiled code

The new maps are better in the sense that they include the dark gas component derived as the residual dust map after linearly fitting the hi and co maps to the SFD E(B-V) dust map. The problem is that the dark gas correction is then assumed to be distributed similarly to the HI map and used to scale the HI annuli up or down, depending on the residuals.

The maps were released because we published the results in a paper. The co maps are identical apart from the resolution degradation (which is minor) and the different annuli (which is major). The release of the co map has not been discussed, but it makes little sense to have incompatible maps in the data package. I will bring this up with the other developers.

Author:  gudlaugu [ Sun Mar 01, 2015 4:18 am ]
Post subject:  Re: Difference results from webrun and locally compiled code

Sorry for the late reply. The CO map used in 1202.4039 is now available in the download section.

Author:  ranlu [ Mon Mar 09, 2015 3:18 pm ]
Post subject:  Re: Difference results from webrun and locally compiled code

gudlaugu wrote:
Sorry for the late reply. The CO map used in 1202.4039 is now available in the download section.

That is great news. Thank you very much.

Page 1 of 1 All times are UTC - 8 hours [ DST ]
Powered by phpBB® Forum Software © phpBB Group
https://www.phpbb.com/