0
When executing a C code I need to execute every 30ms a set of calculations that transforms an image and that is taking up to 100ms which forces me to discard 3 of every 4 frames that I get.
How can I guarantee that the CPU clock frequency is at maximum?
I’m using a Cortex-A53 that runs from 400Mhz to 1400Mhz.
Below is the result of the command cat /proc/cpuinfo
Processor : ARMv7 Processor rev 3 (v7l)
processor : 0
BogoMIPS : 1594.36
processor : 1
BogoMIPS : 1594.36
processor : 2
BogoMIPS : 1594.36
processor : 3
BogoMIPS : 1594.36
processor : 4
BogoMIPS : 1601.53
processor : 5
BogoMIPS : 1594.36
processor : 6
BogoMIPS : 1594.36
processor : 7
BogoMIPS : 1594.36
I even tried to use the code suggested in this link to know how fast the processor, but I’m not able to compile it because I don’t think the correct includes:
https://superuser.com/questions/406141/how-to-get-an-arm-cpu-clock-speed-in-linux
What would be the correct code to get the current CPU clock frequency and how to adjust at runtime to use the maximum?
I’m using Linux, with C++ in GCC 4.9, in a Cortex-A53.
It may not seem like it but the question has to do with programming.
– Maniero
Hello @bigown, I need to make some adjustment?
– Delfino
As information I was able to manually make adjustments using the cpufreq-set tool changing the frequency maintenance strategy to "performance", but even so what I want is for my own code to be able to request maximum processor performance.
– Delfino
The question isn’t a ten, but I think we can answer it. The problem that they voted to close is that the person reads "processor" and thinks it’s hardware :) But it’s clear that it’s not.
– Maniero
Thanks for the comment, I will try to improve it. Any adjustment is welcome.
– Delfino