×

INDI Library v2.0.7 is Released (01 Apr 2024)

Bi-monthly release with minor bug fixes and improvements

Read More...

New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

  • Posts: 185
  • Thank you received: 28

Replied by Richard Beck on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

I missed the rename. I ran /usr/bin/SexySolverTester from the command line and got the correct version. I'll fix my desktop icon to point to the correct executable.

I plan to work on some blind solves with jpegs for input. I'll post those results later today.
3 years 11 months ago #53888

Please Log in or Create an account to join the conversation.

  • Posts: 2877
  • Thank you received: 812

Replied by Rob Lancaster on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

Thank you very much, please note what I said in the post before yours. The Combo Box for the Selector and Solver are now separate from the options on the left. If you want to use the options on the left, you need to select "Unsaved Options" for the Sextracting or Solving options profile as appropriate. Or you can edit a profile, save it, and select to use your new profile for Sextracting or Solving

Also, we found a couple of issues with a couple of the profiles' settings in solving one of Han's images that had many stars near the saturation limit. I will revise these profiles later today, but I haven't done so yet. The main issue was SaturationLimit. It isn't really needed for solving (I don't think?), and it removed too many good stars.
3 years 11 months ago #53889

Please Log in or Create an account to join the conversation.

  • Posts: 185
  • Thank you received: 28

Replied by Richard Beck on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

Rob,

I thought I had run the Internal Solver with the ParallelSolving option several times, but the last couple of times it crashed. For what it's worth, I have attached the crash report.

File Attachment:

File Name: SexySolver....txt.zip
File Size:4 KB


Richard
3 years 11 months ago #53892
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 2877
  • Thank you received: 812

Replied by Rob Lancaster on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

I just updated the options profiles with the downsampling and saturation limit on GIT. You might want to git pull and build again to get a better solve.

I can't replicate your crash. It looks like it happened due to a logging issue. Maybe try changing the log setting to log none or maybe log verb and see if either have an effect on the crash?
3 years 11 months ago #53894

Please Log in or Create an account to join the conversation.

  • Posts: 2877
  • Thank you received: 812

Replied by Rob Lancaster on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

I just fixed a bug with star positions when one profile is used for sextracting and a different one is used for solving. I also added some edit buttons so hopefully the profile lists are less confusing. And I made an automatic option for the parallel algorithms and put it in the FastSolving profiles
3 years 11 months ago #53909

Please Log in or Create an account to join the conversation.

  • Posts: 185
  • Thank you received: 28

Replied by Richard Beck on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

Rob,

I ran a bunch of trials between your last two changes. I haven't analyzed the results, but it seems the "FastSolving" profile is nearly optimum for both the internal sexysolver and astrometry.net.

File Attachment:

File Name: Sexy_Solve...4UTC.zip
File Size:6 KB


I didn't run ASTAP solutions because an early run on blind solves (with jpegs) ASTAP failed. There was a message about atk-bridge missing; however, it was installed and I got the same message after installing the -dev package. A "locate atk-bridge" gives the following:
/usr/include/at-spi2-atk/2.0/atk-bridge.h
/usr/lib/x86_64-linux-gnu/libatk-bridge-2.0.so
/usr/lib/x86_64-linux-gnu/libatk-bridge-2.0.so.0
/usr/lib/x86_64-linux-gnu/libatk-bridge-2.0.so.0.0.0
/usr/lib/x86_64-linux-gnu/pkgconfig/atk-bridge-2.0.pc
/usr/share/doc/libatk-bridge2.0-0
/usr/share/doc/libatk-bridge2.0-dev
/usr/share/doc/libatk-bridge2.0-0/changelog.Debian.gz
/usr/share/doc/libatk-bridge2.0-0/copyright
/usr/share/doc/libatk-bridge2.0-dev/changelog.Debian.gz
/usr/share/doc/libatk-bridge2.0-dev/copyright
/var/lib/dpkg/info/libatk-bridge2.0-0:amd64.list
/var/lib/dpkg/info/libatk-bridge2.0-0:amd64.md5sums
/var/lib/dpkg/info/libatk-bridge2.0-0:amd64.shlibs
/var/lib/dpkg/info/libatk-bridge2.0-0:amd64.symbols
/var/lib/dpkg/info/libatk-bridge2.0-0:amd64.triggers
/var/lib/dpkg/info/libatk-bridge2.0-dev:amd64.list
/var/lib/dpkg/info/libatk-bridge2.0-dev:amd64.md5sums
3 years 11 months ago #53912
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 2877
  • Thank you received: 812

Replied by Rob Lancaster on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

That sounds good, because in my last change, I modified the Parallel solving profiles so that they will automatically do multithreaded solves if it thinks it would help, but if not, then it will default back to the fast solving options. Basically it is a multithreaded algorithm auto setting. So now maybe those will be better all around, because there are times when they REALLY help the performance, but other times when they don't.
3 years 11 months ago #53913

Please Log in or Create an account to join the conversation.

  • Posts: 2877
  • Thank you received: 812

Replied by Rob Lancaster on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

Were you able to reproduce the crash you reported earlier?
3 years 11 months ago #53915

Please Log in or Create an account to join the conversation.

  • Posts: 185
  • Thank you received: 28

Replied by Richard Beck on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

Whatever the default logging was, SexySolverTester would crash. Turning off logging or changing to verbose logging took care of it.
3 years 11 months ago #53916

Please Log in or Create an account to join the conversation.

  • Posts: 2877
  • Thank you received: 812

Replied by Rob Lancaster on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

Ok, I think I can solve it. I will look into it. I think it is an issue only on Linux. Neither windows or Mac had an issue, but sometimes I saw the same issue on another Linux system. It is related to initialization of the logger.

Thanks for testing!
3 years 11 months ago #53919

Please Log in or Create an account to join the conversation.

  • Posts: 2877
  • Thank you received: 812

Replied by Rob Lancaster on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed


Han, I think I fixed most of these things today. #1 I think I fixed by improving the options profile and making the interface clearer. #3 I think I fixed this for the external solvers.

For #2 though, I am confused since I think it does give a warning?



It should be noted that this message prints because my SexySolver algorithm determined it could not run Sextractor since it doesn't exist. It did not try to run the command, but instead returned an error and printed this message. Is this not the warning you are looking for?
3 years 11 months ago #53922
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 2877
  • Thank you received: 812

Replied by Rob Lancaster on topic New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed

One more question Han.

Is there a version of ASTAP released somewhere so that I can try implementing INT SEP EXT ASTAP?
3 years 11 months ago #53923

Please Log in or Create an account to join the conversation.

Time to create page: 0.420 seconds