Maybe of interest, a practical performance test between the Raspberry Pi 3 and Pi 4
Both system where running Raspbian Buster. The camera an ASI1600-MM Cool was connected using an USB 3.0 cable. The imaging program was CCDCiel and ASTAP was installed as the solver as described
. PHD2 was installed for guiding but not used in this test.
The results show that the Pi 4 is in average twice as fast as the Pi 3. A lot of time is lost by downloading and saving the image. A class 10 memory card was installed.
The time lost is the time between the exposures for downloading the image via USB and saving it to the memory card. So in the Pi 4 if you operate the ASI1600 camera in bin 2x2 mode there is 2.5 seconds lost between the exposures. That's pretty good for deep sky imaging.
The solving time is the time required for plate solving the image.
Nice work. I was expecting better performance though. Is it possible the Pi 4 was being thermal throttled?
INDI/KStars on Raspberry Pi 4, 4gb
Raspbian Buster with AstroPi3 script configuration
Skywatcher HEQ5 Pro Mount
Canon 600D Camera
Orion SSAG/ASI120mm @280mm Guide Scope
Adafruit Motor Hat shield
Adafruit GPS Module
Generic Bluetooth Joystick.
Startech 7 port powered USB Hub.
I don't think the Pi 4 was throttled. This setup was all local in the Pi 3 and Pi 4. No download or INDI communication via the network
For local saving, the memory card has a significant influence. A class 10 can do 10 mbit/second. So the normal sequence is something like:
1) start exposure
2) download result from camera
3) save result
4) start next exposure
Patrick is already working on a improved CCDCiel version where step 3) and 4) are run in parallel. So a multitask operation. This would reduce the lost time (between exposures) to the image transfer time from camera to memory only. First indications are that the lost time will be half or less.
This are the values for my desktop under win7:
win7 desktop, 4635 x 3520 image, lost time between images 9.3 sec
win7 desktop, 2328 x 1760 image, lost time between images 3.3 sec
So the Pi4 is faster then my old desktop!
An alpha version of CCDCiel indicates that the 9.3 sec could reduce to 3.6 seconds and 3.3 seconds to 1.3 seconds. The same improvement could apply for the Pi 4, but it is early in the development.
First I tried a ram disk in memory, but that not helped a lot. Selecting in the driver option save to ram disk (SD card) rather then via INDI network communication helped a little.
But in the latest alpha version of CCDCiel the problem is solved and only the download speed from the camera to RAM is relevant (USB 3). Saving is done while the next exposure is running. The lost time between exposure is half of the previous values or less. This improvement will be beneficial for all CCDCiel releases. At the moment I can only test on the Pi 3 since my Pi 4 is broken.