Linux support for ARM big.LITTLE

Posted Feb 23, 2012 9:44 UTC (Thu) by slashdot (guest, #22014)
Parent article: Linux support for ARM big.LITTLE

How about using the cpufreq algorithms, but applying them to the union of all CPUs in the system?

Linux support for ARM big.LITTLE

Posted Feb 23, 2012 9:47 UTC (Thu) by slashdot (guest, #22014) [Link] (2 responses)

(where obviously the "power states" would be A7-only, A15-only and A7+A15).

Linux support for ARM big.LITTLE

Posted Feb 23, 2012 9:55 UTC (Thu) by slashdot (guest, #22014) [Link] (1 responses)

And if you can really power up/down each core independently with power savings, then have (n + 1) * (n + 1) power states consisting of all #A7 ad #A15 combinations, sorted either by increasing total energy consumption or increasing total bogomips.

Linux support for ARM big.LITTLE

Posted Feb 23, 2012 19:49 UTC (Thu) by dlang (guest, #313) [Link]

there are corner cases that need to be figured out even with this model

1. if you have a CPU hog, how well/poorly does it work if you have one core that is 2x (or 10x) the power of a different core? can your cpu hog get stuck on a slow core?

2. how do you deal with software that will adapt to not having enough processor

3. what about software that's constrained by memory bandwidth, not cpu cycles, it may be just as fast on a slow core as a fast one, but use 100% cpu in both cases (since CPU utilisation doesn't reflect time spent waiting for memory access)