hyc

Patch looks ok. I wonder if we should instead make the instructions to set a shell variable for the version instead, so we only need to update the doc in...

Just compiling with `-march=native` is supposed to tailor the generated code to the current machine's CPU. There should not be any other action needed. If `-march=native -Ofast` generates an unusable...

> * losing the ability to soft-fork the transaction format This bothers me a bit, but really, when have we soft-forked anything? So far all changes have been hard forks.

A lot of these single board computers have issues with weak power supplies. I'd look into that first of all.

Is your current power supply one that was originally bundled with the rockpro64? At 60W it should be fine already; you shouldn't need anywhere near 60W unless you've got a...

I actually meant checking the power rails on the RockPro64 while the power supply is plugged in and the rockpro is powered up and running. You should be able to...

I've had zero problems with my rockpro64. Had no idea any memory instabilities were a common thing for them.

> based on this thread: That thread is for the older Rock64, not for RockPro64. Completely different chip. > when I run this in dir: > /sys/class/devfreq/ff9a0000.gpu$ The GPU should...

At this point the experiment seems pointless. Do a full sync on a PC, then copy the blockchain to the RockPro64 and see if it can stay sync'd without crashing.

> The overall comment I have is about changing the style from a ~80 character break to an unterminated one. Does anyone else care? The markdown formatter doesn't care, the...