What is PVTM? Or why your Rockchip RK3588 CPU may not reach 2.4 GHz

cnx-software.com

38 points·zdw··6 comments

I love the comments debating over what chip specification should be advertised... if a chip is stable at 2.2 GHz, and many are unlikely to perform at 2.4 GHz, it should just be advertised as a 2.2 GHz chip (despite many being able to perform better).

The BCM2711 in the Raspberry Pi 4 model B (even the older B0 revision) was clocked at 1.5 GHz by default to save on heat, and for stability, even though most (in my testing, all I've ever used) can run at 1.8 GHz without issue in most environments, and 2.0 or more with better active cooling.

Hi Jeff! Just want to say thanks for all the great stuff you put out for Pi. It really helped me get mine going (after keeping it in drawer for 6 months).

Ah yes, advertising turbo clocks as if they're base clocks. Sadly this is neither the first nor last time this will happen.

This smells more like an integration issue to me. The chip is essentially throttling because it’s reaching a high temperature or low voltage threshold. It could just be a bad power supply or the hot ambient temperature (28c ~80f) mixed with inadequate cooling.

The same thing happens on a lot of mobile devices as well. It's the reason that mobile devices generally don't get advertised with clock speeds. E.g., see: https://dl.acm.org/doi/abs/10.1145/3032970.3032982

Unlike desktops, where CPUs from different quality bins are sold at different price points (with different clock speeds), mobile CPUs are simply run at different voltages to compensate for quality differences. The resulting differences in power dissipation end up creating different temperature profiles, depending on which bin you happen to get.

Isn't the RK3588 an 8nm part? Seems odd that it'd have this sort of issue at that frequency, no?

Reply on news.ycombinator.com