15th Feb 2024: Astro Pixel Processor 2.0.0-beta29 released - macOS native File Chooser, macOS CMD-Q fixed, read-only Fits on network fixed and other bug fixes
7th December 2023: added payment option Alipay to purchase Astro Pixel Processor from China, Hong Kong, Macau, Taiwan, Korea, Japan and other countries where Alipay is used.
Last night I compared APP on two laptops both running Windows 10 Pro with SSD drives
1. : Intel 6th Gen i7-6700HQ up to 3.50GHz with 24GB DDR4 RAM (4 cores with hyper threading = 8 threads)
2. Intel 4th Gen i7-4700MQ up to 2.4GHz with 8GB DDR3 RAM (4 cores with hyper threading = 8 threads)
- Lights x 71
- Flats x 50
- Dark - Flats x 50
- Master Bias x 1
- Master Dark x 1
- Bad Pixel Map x 1
The only non default settings were under 6) Integrate
- Weights=quality
- LNC Degree=1st Degree
- Outlier Rejection=LN Winsor Clip
The first laptop took 30 minutes to run through all the processes
The second laptop took 33 minutes to run trough all the processes
I was expecting a bigger improvement in speed. What should I be looking at to get a better improvement in processing time?
Anyone would like to make some recommendations based on personal experiences with hardware and APP?
Integration speed will and processes like LNC & MBB will benefit from faster drives. So a M2 SSD will be much faster than a SATA SSD which is limited by the SATA pipeline.
Most calculation are done using multiple cores, so the more cores the better. 32 threads does translate to 4x the speed in most calculations when compared to a 8thread machine.
I would expect the Intel 6th Gen i7-6700HQ to outrun the Intel 4th Gen i7-4700MQ though... perhaps the IO on the harddrive was the biggest limiting factor in your test.
So in my own experience, harddrive read/write speed has a big influence and the amount of cores. In future APP versions, your videocard will become more important, since I will try to run more algortihms on the GPU for more speed increase.
Kind regards,
Mabula
In future APP versions, your videocard will become more important, since I will try to run more algortihms on the GPU for more speed increase.
Mabula - if you implement this, can you give us an option to disable gpu processing? I'm super intrigued with the remote linux server build strategy for an APP processing server, but don't want to have a big GPU card on that remote server. So if you implement this without deterministic logic whether a GPU on the machine can support it, could be problematic ya?
