Mar 28 2026 APP 2.0.0-beta40 will be released in 7 days.
It did take a long time to have the work finished on this and it will have a major performance boost of 30-50% over 2.0.0-beta39 from calibration to integration. We extensively optimized many critical parts of APP. All has been tested to guarantee correct optimizations. Drizzle and image resampling is much faster for instance, those modules have been completely rewritten. Much less memory usage. LNC 2.0 will be released which works much better and faster than LNC in it's current state. And more, all will be added to the release notes in the coming weeks...
Update on the 2.0.0 release & the full manual
We are getting close to the 2.0.0 stable release and the full manual. The manual will soon become available on the website and also in PDF format. Both versions will be identical and once released, will start to follow the APP release cycle and thus will stay up-to-date to the latest APP version.
Once 2.0.0 is released, the price for APP will increase. Owner's license holders will not need to pay an upgrade fee to use 2.0.0, neither do Renter's license holders.
Hi,
I'm looking for tricks to speed performance as I stack a ton of short exposures.
I have two systems:
iMac: Four core intel 24GB/Radeon M395
Windows system: 8 core AMD Ryzen 9 5900HX with 16GB RAM with Nvidia 3060
Both have high speed SSD drives.
I tend to stack many many exposures of 26Megapixel ASI2600MC OSC data (I shoot short exposures for reasons I won't get into here). So think 500-5000 images... a ton. (Yes, I'm playing with capturing some images in bin2 for speed).
How much would increasing the amount of RAM to 64GB improve my performance? I notice it allocates nearly all the RAM if I let it, so I was wondering.
Finally, with respect to the CPU/GPU, clearly the GPU isn't used for stacking (sad!), but given I have an SSD, is it pretty much CPU core limited? (i.e. go from 8 to 16, you get nearly 2x, go from 16 to 32 core thread ripper, get another nearly 2x jump?).
Regards,
Steven
APP currently uses the CPU for most parts, GPU support is something we intend to implement, but require a lot of study and isn't the top priority at this moment. More ram will help, but more cores definitely as well, APP is able to use as many as you throw at it (unless it simply doesn't need the amount of cores you have, then it will use what it can).
Adding to Vincent's answer, what version of APP are you using? If one of the 2.0.0 betas then note that that will automatically allocate all system memory but 2Gb. This may give the impression that APP uses nearly all memory when in fact it may not actually require that.
Adding to Vincent's answer, what version of APP are you using? If one of the 2.0.0 betas then note that that will automatically allocate all system memory but 2Gb. This may give the impression that APP uses nearly all memory when in fact it may not actually require that.
Ahh, this is good to know, so that makes it more difficult for me to easily see it bumping into RAM constraints as it's unclear how much of the allocated RAM is really utilized. So I can throw more memory at it, but it's unclear of how much I would gain. However, more CPUs are a clear and fairly proportional improvement.
For RAM, I was unsure if there was a threshold where some intermediate results could be saved in RAM and I would then get a speed bump.
Also, how is the performance on Apple's new M1 Cores relative to the Intel cores? Eventually I'll be upgrading my iMac and I hope they will come out with an iMac Pro with a >10 core M2 type chip.
@readyjetty To clarify what I added to Vincent's answer, APP will run poorly if there is not enough RAM. I see you currently use 16 Gb which is sufficient but more may be required if you increase the number of cores. Both are coupled of course. I merely wanted to state how APP allocates memory and by no means wanted to indicate that APP over-allocates!
Hi @Steven Miller,
Regarding the MAC performance, I can tell you, it is quite a jump: I had a 27" imac late 2015 w/ 4GHz i7-6700 & 24 GB RAM before and recently upgraded to a Mac Studio w/ M1 Max & 64 GB RAM. The M1 Max is 10-core and all cores get nicely loaded by APP, especially during Analyze Stars and Integration. Also, the increased RAM, as well as the great memory bandwidth and the unified memory architecture really help a lot.
I used to stack a few hundred 24 Mpix shots as well and had the process usually run overnight (8-10 hrs)with the i7. It now completes within 2-3 hours (much slower if I use LNC which is still single core only, hence it only scales by 1.7 x single core performance or so) 🙁
Some calculations had become possible with M1 Max only e.g. higher degree LNC). As APP is CPU only, you might save money using the binned 24 core GPU model (This still renders pretty fast and is usually never loaded 100%, even with StarTools, which has some GPU optimized processes).
I hope APP will get more optimized going forward, at least LNC. On Mac it is possible to just address the Core ML layer which distributes the workload between GPU and NPU cores on its own. That would give a huge extra boost. We have decent hardware out there, the bottleneck is software optimization.
Cheers,
Jochen
LNC is definitely worked on, so I hope that will get better indeed. 😉
Thanks for the info, that’s good news on the performance boost on the M1 Macs.
I will say that stacking performance is a real driver for me as I’m now trying to reduce data as I shoot so many short exposures. I either bin2, use a subset of my full camera frame, or just point on collecting “even more data”. So either I invest in a traditional fully guided system, or computational limits will hold me back. I’m pining for the days when Moore’s law was not dying: A doubling of computer speed and storage size every 18 months was a luxury when it lasted.
Now it’s up to software and algorithm optimizers to bring the big performance gains.
