Quote from: fellestreum on June 06, 2018, 07:50:33 PM

Thanks MagicSmoker for the guide, I think people will find that very useful.

I know it isn't much, but sometimes our glorious Captain forgets about the skill level of us mere mortals when it comes to this sort of thing... Grin

Quote from: fellestreum on June 06, 2018, 07:50:33 PM

Note that the Windows way of syncing to a timeserver could cause problems. If you find that even after checking to make sure that Windows is set to update the clock, and has (it claims) updated the clock, you still get complaints about the clock being out of sync from the daemon, come back here and we will try to solve that problem.

I've lost sync twice in less than 12 hours so I already resorted to changing the timeserver to pool.ntp.org as suggested on the Slack channel, but I really don't think this an acceptable solution for the general public. Yes, it is fairly easy to make this change, but it also the kind of thing that M$ could change back at any time in one of their infamous forced updates (you know, the kind you can't block no matter what services you stop). For example, the 1803 update now makes Bing on the Edge browser the all-but-unchangeable means of searching via the desktop; I was none too pleased about that, as I despise both Bing and Edge.

Quote from: fellestreum on June 06, 2018, 08:16:04 PM

Quote from: MagicSmoker on June 06, 2018, 01:24:30 PM

Quick Testnet How-To For Windows

Replace with your address and <# of threads> to a maximum of the amount of L3 cache divided by 2; e.g. - a Ryzen 5 1600 has 16MB L3 cache so max number of threads is 8 (but to preserve desktop interactivity set to 1 or 2 less threads than maximum).

I'm not sure this is good advice for all cases. I have a 44 core machine with 55 MB of L3 cache. Mining with 27 threads (55 / 2) I get about 62% of the hashrate I get by mining with 44 threads. (I have turned hyperthreading off, so have not tried with 88 threads, but I doubt that would see better performance, and it could well be worse.)

Yeah, I fully expect to have to tweak this, but until we get an actual explanation of how the new PoW algo works I fell back on the usual recommendation for straight CryptoNight, as that seems to be working for /most/ of the people on Slack.

That said, your CPU is distinctly unusual and so what works/doesn't work for you might not apply to those of us with rather more pedestrian processors.