I wanted to come up with a quick way to test GPIO sampling speed. The documentation for the gpio-led dtoverlay shows that there is a "gpio" trigger that should effectively mirror an input pin state to an output pin.
I was using the documentation from here: https://github.com/raspberrypi/firmware ... ays/README
I tried this on a Pi4 (and a Pi2), as those were what I had easily available for testing. Neither seems to have that "gpio" trigger as an option in
cat /sys/class/leds/myled1/trigger
Is there any other easy way that I can test how fast the GPIO can react to a pin change input trigger?
I was actually considering writing a kernel driver to sample one GPIO on the rising (or falling edge of) another GPIO input. This would be something like a clock input and a serial data input (like synchronized serial data). My hope was to test the speed before I put effort into writing any code.
Is there any chance that I could do this reliably if my clock input to the Pi is 20MHz (maybe 25MHz)?
If not, what would be the fastest I might expect to do this?
I know that I said I was using a Pi4, but I could likely move to a Pi5 if it could do the rates I'm talking about.
I was using the documentation from here: https://github.com/raspberrypi/firmware ... ays/README
I tried this on a Pi4 (and a Pi2), as those were what I had easily available for testing. Neither seems to have that "gpio" trigger as an option in
cat /sys/class/leds/myled1/trigger
Is there any other easy way that I can test how fast the GPIO can react to a pin change input trigger?
I was actually considering writing a kernel driver to sample one GPIO on the rising (or falling edge of) another GPIO input. This would be something like a clock input and a serial data input (like synchronized serial data). My hope was to test the speed before I put effort into writing any code.
Is there any chance that I could do this reliably if my clock input to the Pi is 20MHz (maybe 25MHz)?
If not, what would be the fastest I might expect to do this?
I know that I said I was using a Pi4, but I could likely move to a Pi5 if it could do the rates I'm talking about.
Statistics: Posted by Flavor — Wed Jan 29, 2025 12:07 am — Replies 1 — Views 25