| Apple MacBook Pro (14-inch, M1 Pro, 2021) | 3036/2064 MHz | 5.18 | Gentoo Base System release 2.8 arm64/aarch64 | 43800 | 620960 | 1064450 | 27110 | 71910 | 48.28 | http://ix.io/443N

ThomasKaiser / sbc-bench

Simple benchmark for single board computers

BSD 3-Clause "New" or "Revised" License

658 stars 78 forks source link

| Apple MacBook Pro (14-inch, M1 Pro, 2021) | 3036/2064 MHz | 5.18 | Gentoo Base System release 2.8 arm64/aarch64 | 43800 | 620960 | 1064450 | 27110 | 71910 | 48.28 | http://ix.io/443N | #47

Closed chadmed closed 2 years ago

chadmed commented 2 years ago

Some things to consider:

The memory controller supports two pstates, but we currently do not activate the higher performance one
The top pstates for the Firestorm cores are disabled in the devicetree pending changes to PSCI. Firestorm cores should hit 3.2GHz on these machines.

http://ix.io/443N

ThomasKaiser commented 2 years ago

Thank you!

You've been too fast for me since I just added the new core definitions after you started and as such your benchmark run didn't catch the full set of cluster tests MODE=extensive usually runs. But this doesn't matter now. In case you're enabling the higher pstates revisiting this would be great.

Keep up the great work!

ThomasKaiser commented 2 years ago

@chadmed out of curiosity: is there a specific reason to choose CONFIG_HZ_1000 in your kernel over e.g. CONFIG_HZ=300?

chadmed commented 2 years ago

The marginally lower latency for desktop/multimedia use outweighs the increased power draw, especially considering we have no real power management yet and are relying on software rendering anyway so battery life sucks regardless.