astap program / Forum / General discussions ASTAP: Probleme mit der aktuellen Version 2025.03.07

fedorauser - 2025-03-10

Good morning, I just wanted to report some peculiarities in this version that are causing problems.

the DEB Ubuntu version has massive problems in the GUI and throws numerous error messages.
There are also some options overlaid by other GUI parts.

the MacOS Apple_silicon M1 version is 20 times slower compared to old versions and I can't find an explanation for this.
I have also completely cleaned the system including the databases and still had no success.
I am currently back on astap_M1 v2024.07.26

CS

Translated with DeepL.com (free version)

Bildschirmfoto vom 2025-03-10 07-36-29.png
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2025-03-10

Thats worrying.

One strange thing is the missing canberra-gtk-module in your system. Which Ubuntu verion are you using?

If it is an older Ubuntu version can you try the version compiled for older Linux systems:

https://sourceforge.net/projects/astap-program/files/linux_installer/astap_amd64_older_linux.deb/download

Han

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

fedorauser - 2025-03-10

I use the 24.04.2 LTS ... The Canbera GTK Problem has now been over a decade old and can be found in all distributions and also affects other programs as I could see.

Do this error messages have to do with the peculiarities of the GUI that some buttons and functions are overlaid by others?

Last edit: fedorauser 2025-03-10

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2025-03-10

I have run ASTAP in my Lubuntu 24 and 22 virtual machine. The Canbera GTK Problem does not show up but the other messages do. They do not see to do any harm.

The latest ASTAP should be about 1.8 faster not slower. The big difference is multiprocessor/theading in the latest version.

To understand if it is the GTK2 interface or something else, can you try the QT5 version?

This one version only available in an archive.
https://sourceforge.net/projects/astap-program/files/linux_installer/astap_amd64_qt5.tar.gz/download

To make it run your have to install the QT5 interface as follows:

sudo apt install libqt5pas-dev

This version still has a problem that the check marks are not sorted. Does this version stacks at normal processing speed?

Cheers, Han

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

fedorauser - 2025-03-10

ich werde mich daran versuchen...
was es mit der Mac_silicon Version auf sich hat bin ich auch etwas Ratlos

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2025-03-10

You could unpack the QT5 version anywhere at you home directory. There is no need to move it to /opt/astap.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

fedorauser - 2025-03-10

I have also tested it in a VM with Ubuntu 24.10. The error persists.

the QT5 variant does not show this error but takes 10 times as long to start and displays this message in the terminal.
[WARNING] GetPosition called without handle for PairSplitter1(TPairSplitter)

Last edit: fedorauser 2025-03-10

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2025-03-10

You have two problems. I need more information to understand the probem

Error mesage in Linux. Do they slow down the process either the normal version or the QT5 version?

The MacOS Apple_silicon M1 version is 20 times slower. Is this in analyse or stacking? Can you attach a full log of stacking. Using the log times I possibly locate the problem area. So press analyse and then stack using sigma clip stacking method. Then copy the log in a post or or attachs as txt file.

I will continue tonight. Probalby it is wise to release a other version just before the threaded version was issued.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

fedorauser - 2025-03-10

in Linux Ubuntu it doesn't slow down, 8 cores are utilized and that's great.
Initially, only these error messages were displayed in the console, otherwise there were no problems and it runs really fast. Only the QT5 version takes about 5 seconds to start, the normal version starts immediately.

Regarding the current Apple Silicon, Apple has changed something again and the execution of

codessign - -force -s -/applications/astap.app/contents/macos/astap

does not work .
codesign - -force -s -/applications/astap.app/contents/macos/astap

returns as feedback

invalid argument “rce”.

The execution of Astap is also extremely slow when analyzing the data and not only when stacking.

I'll clean up everything again and install it again in the DS80 database and get back to you.

Translated with DeepL.com (free version)

Last edit: fedorauser 2025-03-10

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

fedorauser - 2025-03-10

Short feedback, no improvement experienced.
Analyze per frame approx. 2 seconds and when stacking 10sec/frame with a load of one CPU core.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2025-03-10

Thanks for testing. I have recreated version 2027.02.17.

https://sourceforge.net/projects/astap-program/files/macOS%20installer/older/

Can you check if this version is fast as expected? Could you both check the X86_64 and M1 version? I have to know if the problem is really caused by threading. I have also upgraded the compiler. That could be an other cause.

I will do now some speed tests on my older Mac with an Intel processor.

Then a special request, could you allow me remotely to compile and debug the program on your M processor Mac?

Cheers, Han

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2025-03-10

The threaded version is on my Intel Mac not faster. The reason it sees only one processor. This is not an explanation for your problem but at least something.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

For the Apple Silicon variant
The analysis of 361 Full HD Lights took place in 96SEK with up to four kernels

The complete stack including analysis needs 329sec absolutely sensational

22:31:07  █ █ █  Saving result 361 as /Users/martin/Astro/M 42_mosaic/M 42_mosaic_sub/lights/M 42, 2025-01-29, 359x10L, S50_852b0edf, (IRCUT), Seestar S50_stacked.fits
Finished in 329 sec. The FITS header contains a detailed history.

HISTORY 1   Stacking method SIGMA CLIP AVERAGE                                  
HISTORY 2   De-mosaic bayer pattern used GRBG                                   
HISTORY 3   Colour conversion: AstroC, colour for saturated stars. interpolation
SET-TEMP=                   20 / Average set temperature used for luminance.    
LUM_EXP =                   10 / Average luminance exposure time.               
LUM_CNT =                  359 / Luminance images combined.                     
LUM_DARK=                    0 / Darks used for luminance.                      
LUM_FLAT=                    0 / Flats used for luminance.                      
LUM_BIAS=                    0 / Flat-darks used for luminance.                 
DATAMIN =                    0 / Minimum data value                             
DATAMAX =                65524 / Maximum data value                             
CBLACK  =                  752 / Black point used for displaying image.         
CWHITE  =                 2568 / White point used for displaying the image.

Last edit: fedorauser 2025-03-10

fedorauser - 2025-03-10

Vielen Dank für die spontane Unterstützung.

CS

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

fedorauser - 2025-03-11

I have tested the Linux version in a VM with Linux Mint and in this case the GUI fits in terms of the layout.
It looks like it is due to Ubuntu and the GNOME version.
And also the error messages do not appear in the logs.

CS

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2025-03-11

I have a solution for the Mac. I will release in an hour.

Also the Linux version is too slow. All related to CPU count detection

Cheers, Han

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

han.k - 2025-03-11

It should work now in the MacOS version 2025.03.11. Can you check? There where two problems.

1) The threads where running on a single processor because the number of available CPU detections did not work for MacOS.
2) I did by mistake put the X86_64 in the M package.

The Linux warnings I will look into somewhere in the future.

Cheers, Han

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

fedorauser - 2025-03-11

das Ergebnis für MacOS 15.3.1 M1 MacBook Air 16 GB UMA

~~~
17:56:42 █ █ █ Saving result 361 as /Users/martin/Astro/M 42_mosaic/M 42_mosaic_sub/lights/M 42, 2025-01-29, 359x10L, S50_852b0edf, (IRCUT), Seestar S50_stacked.fits
Finished in 393 sec. The FITS header contains a detailed history.
~~~

It was a bit faster yesterday, but I have to use yesterday's config again and compare it again.

All in all, the computing power behind it is impressive - thank you for this innovation

Last edit: fedorauser 2025-03-11

Bildschirmfoto 2025-03-11 um 18.25.01.png

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Probleme mit der aktuellen Version 2025.03.07

Stacking program, astrometric (plate) solver, fits file viewer

Forums

Help

Probleme mit der aktuellen Version 2025.03.07

Probleme mit der aktuellen Version 2025.03.07

Stacking program, astrometric (plate) solver, fits file viewer

Forums

Help

Probleme mit der aktuellen Version 2025.03.07 document.SUBSCRIPTION_OPTIONS = { "thing": "topic", "subscribed": false, "url": "subscribe", "icon": { "css": "fa fa-envelope-o" } };

Probleme mit der aktuellen Version 2025.03.07