Quote from: UnclWish on May 20, 2020, 08:58:19 AM

Hi, kerney666!
No questions about fees. But can you add in your miner info about devfee mining when it goes? At least start mining devfee and stop.

Hi!

Well it's not that simple, there's just no point really. We don't run mining the same way as the traditional switching miners do, we always mine user and dev in parallel, all the time, for all algos and situations where it's possible which is almost all. 99 hashes to the user, 1 hash to the dev, 99 hashes to the user, 1 to the dev, and so on. The user pool is always connected, the dev pool is always connected. We can print "User and dev fee now mining" at the start and that would be it

.

We built it this way for a much smoother user experience: continuous hashrate for the user with no drops in pool avg, much lower risk for pools being angry that you stop submitting shares for a while, the pool diff stays constant and doesn't fluctuate with dev fee sessions, absolutely the same continuous gpu load all the time with no short interruptions when switching, etc. Now, if you'd be evil you could argue we built it this way to hide our super massive dev fee hashrate that really is 500% more than we're saying, but yeah, simply not the case...

There are a few notable exceptions though, MTP is one of them. We still try to switch different gpus at different times to keep things smooth for the user and it would honestly just (1) litter the logs with and (2) invite even more hackers that try to pick things apart by logging dev fee start/stop. Moreover, I can count the nr of users we have for MTP on one hand and none of them has complained Grin

.

For miners that don't fully trust the displayed hashrate or that the communicated dev fee is correct, I would argue that the best way to verify poolside/dev fee hashrate is with fake pool hacks, much like the one we built ourselves for ethash, available on my github. It sets a low diff, the miner you want to test submits many many more shares compared to a normal pool, and after maybe 12h you have a very accurate picture of the "poolside" hashrate of the miner you're testing because you'll have submitted 500,000 shares by then and get a much much more accurate data point than looking at a 6h or 24h average on some pool for 1000 submitted shares (which really means nothing tbh).