New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed #54624
For the last image I think there is conflicting information about the dimensions. In ASTAP viewer it's indicated 0.72 x0.72 degrees. However the SExysolver command specifies -fov 0.238979 There must be a keyword value in the header wrong.
I have renewed my index files using the ANSVR downloader. Didn't help. Same errors like:
C:\CraftRoot\build\_\53d1cff2\sexysolver\astrometry\util\starkd.c:266:my_open: Failed to read kdtree from file "C:/Users/h/AppData/Local/cygwin_ansvr/usr/share/astrometry/data/index-4218.fits"
New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed #54668
I got this TIFF file from the Nina program support group. This TIFF format can't be read by Sexysolver. It must be a limitation of the TIFF reader. The file is readable by other programs. I will try to find out the origin.
New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed #54691
I am limited in loading multiple index files (one form of parallel processing) by the 1G of RAM I have in my Rpi 4. I see that I could now boot my Rpi4 from my Sandisk 128 G Extreme USB Flashdrive and it should be faster than from the SD card.
I also found where I can set up a 'swap' file memory on a USB 3 flash drive. I can put a small flashdrive (say 16G) in one of the USB 3 ports of the Rpi4 and have this recognized as the 'swap' file by the system. Would this allow me to run 'parallel' index processing or is that strictly limited to internal RAM? I know internal RAM is accessed faster than the USB 3 flashdrive, but I thought maybe by having the 'parallel' processing capability it might make the solves run faster.
Is this doable and does it require any changes on the part of ASTAP or any of the internal astrometry solvers?
thanks, Just messing around during our week of cloudy nights courtesy of Bertha.
New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed #54714
Yes, it does make it faster if it can load all the index files. You are correct that a swap file would increase the memory available and would allow it to not crash during a solve with all the index files loaded. However, swap files are orders of magnitude slower to read and write to than RAM. I also would not recommend a swap file in the main SD card of a raspberry pi because it will wear out the card much faster. But you did say the swap file would be on an external drive, so that would be fine. BUT, one solution you maybe did not try is ZRAM. I have had great luck in running KStars and other software with ZRAM enabled. It slows down processing a little, but not nearly as much as SWAP files do and it effectively increases the available RAM without sacrificing the longevity of a flash drive..
New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed #54717
Note that currently I have set SexySolver to check available memory and automatically turn off the inParallel option if there is not enough memory on the system to load all the desired index files. Currently, I think the command that I am using just gets the installed system memory, not any swap files. But it could be changed in the future. I don't want it crashing due to memory issues.
New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed #54737
Optimized images to analyze when embedding in Ekos (pixel size, file size for quick analysis, etc.)
Please consider it as well.
When solving with Ekos, I am currently setting Jpeg / downsample on the DSLR and binning / downsample on the astronomical camera to increase the image transfer rate. If you have a frame from live view and you have an astronomical camera, shooting in stream mode will be even faster. (I think ASIAirPRO improves processing in this style.)
New Internal Solver for Mac, Windows, and Linux -- Testing/ Experiments needed #54813
CapnRon wrote: Han,
Is version 0.9.365 still the latest version for the Rpi? that is all I find on your site.
Yes no later versions are released. An Rpi update will be released in a few days. I want to be sure that the versions released on Sourceforge are bug free. The new version will be more tolerant against elongates stars.