Hi,
We have a strange periodic pattern in our waves. We are simply toggling over and over our PIO without any interrupts. But every 1ms we are not toggling for 40us.
Our code is stored on On-Chip memory.
We wonder if :
1) The BRAM needs to be refreshed like an SDRAM
2) The JTAG has some interrupts even when running and not debugging.
Is one of our guesses right or are we missing something else?
Best
Hey!
We also measured strange things for this manip (though we get a frequency around 3 MHz...), and I think it is solely due to aliasing effects, the logic analyzer's frequency being a lot smaller than 2 times system's clock frequency. I guess that my partner and I just won't spend much time trying to interpret those results :-)
Hi all,
That does seem indeed funny. Both your explanations could make sense, depending on your code and if you're using caches, it could also be that you get cache misses, for instance if you have an infinite while loop and for some reason at some code path, some lines get evicted. I can't really know for sure without having a look at the system.
For the context of this lab, you don't need to look much further into this issue, however make sure that you take many measurements and average them together. Outliers are sure to happen, especially if you're using caches.
If you were designing a system that is long-running and in which timings are hard constraints, you would of course need to make sure whether this glitch is due to measurement error, or from the system itself.
Best,
The TAs team
Follow the pulses of EPFL on social networks
© 2023 EPFL, all rights reserved