is it worth trying to add some optimizations from e.g. xmrig into monerod's miner?

I get the feeling the mutex overhead will already kill it so maybe not

<\x> itll be nice

Definitely worth it for changes that don't take much time to port.

I assume there's big bang/buck ones and medium/small bang/buck ones.

i think its worth it. Currently solo miners using the daemon are at a disadvantage compared to running a 3rd party miner against their daemon. would be nice if they were the same

Huge pages + MSR should reduce the difference to a few %

xmrig has randomx_boost.sh to apply MSR mod separately

randomx-benchmark + huge pages + MSR mod = 8220 h/s, xmrig = 8750 h/s, hmm... I though the difference would be smaller

<\x> sech1: tried emailing hwbot again?

<\x> xmrig randomx benchmarks on hwbot plez

not sure how I feel about writing MSRs inside monerod

It could be made a non default option if it's got potential dangers.

it can slow down other parts of monerod

it's better to keep it separate

in that case there doesn't seem to be much to do on this front, in monerod itself

total miners 85000

minexmr.com 32.6

<\x> make the msr thing opt-in

<\x> the tweak affects some things, makes other programs slower

Does adding msr increase the attack surface for an adversary?

msr only works with root privileges so yes