Best hardware and linux setup for stacking?

@mabula-admin, any recommendation from your end? I know you have quite some different hardware setups for APP testing. Is there some advice from your side, please? Thank you and a happy new year!

APP is Java based so largely isn't impacted by Windows, Mac or Linux.

I'm old-skool, so prefer to run Java based apps (like APP) on Mac (BSD-like) or Linux. I use APP on my iMac and Linux.

So generally, all the same rules apply regardless:

- It currently ignores your GPU for processing on all platforms, but will use it for loading your image (via openGL).

- More RAM the better for large integrations

- More CPU Cores/Threads and Core Speed, the better for faster processing as multiple sections of processing occur in parallel across all available/allocated CPUs.

- NVMe is generally preferred for reading/writing but I have NVMe, SSD (via USB-C) and 7200 HDD and there wasn't too many gains vs each other. With Linux + more RAM, you get efficiences with disk buffers and caching that multiple reads of the same files (from cache) on a HDD is comparable to reading directly from NVMe, and therefore once loaded into cache, the performance is the same -- only the initial read or final write will you see differences. We're talking seconds though, not something you'd care about. 

I recently upgraded my Ryzen 5 3600 to a Ryzen 7 5800X and got 2x performance increase with 32Gb Ram. I recently upgraded to 64Gb Ram to accommodate more mosaics (and previous session integration) -- so depending on how large your image processing input files are will determine if you need more Ram.

But as people will point out, even if you run out of RAM on a mosaic, you can always workaround this by processing Mosaic in different pieces (smaller chunks, then integrate these processed chunks rather than the processing all at once).

So in Summary: OS doesn't matter, whichever you prefer. I prefer control of *nix OS, and prefer Linux and have less confidence in control of Windows, but if you are more comfortable, just avoid Win 11 as it's awful. If you have >32Gb RAM, a better CPU will see your biggest gains. HDD vs Solid State -- really not that much difference overall, but these days you'd choose solid state (ssd or NVMe) anyway to make this a moot point. RAM: 32Gb at least, 64Gb+ if you're really taxing APP with large data sets (but as stated you can work around this anyway).

@itarchitectkev, thank you so much for taking the time and sharing your experiences.

This was great Input, especially the hint of doing mosaics in smaller chunks. I had the same thought in the past but no opportunity yet to test. So instead doing a 8x8 I can try doing for 2x2 and add those into another larger 2x2 later to get the same result but with less hardware consumption at a time. I mean, I stack each panel separately first anyway. So it`s not a big difference splitting the combination of the final chunks. I was also not aware that APP is not using GPU for the combination. I always thought it would...

I had also the impression that Linux could be beneficial over Windows. Again, thanks for the great input!

Hi Patrick @artaios and @itarchitectkev,

So instead doing a 8x8 I can try doing for 2x2 and add those into another larger 2x2 later to get the same result but with less hardware consumption at a time.

This is something that I can not recommend, the optical distortion correction will suffer in this workflow, the best results will be delivered when you properly calculate the whole mosaic at once 😉

I agree that the OS does not matter really. The hardware does, and more CPU cores/ Threads will simply have a more faster APP :-), for big mosaics, 64GB is plenty at the moment. More is not needed since we still have an image size limit for big mosaics of 2^31 pixels = 2 gigapixels. We will try to remove this limit going forward of course. Wiht 64 GB, you can already create huge mosaics. Finally, the faster SSD drive that you can afford will help as well.

Mabula