I'm wondering if this is my issue.

Two separate machines, windows 10 1709. CMD remains open and reports healthy hashrate of previous but GPU load will randomly go to 0 as if something crashed and stay there. Pool reports no hashrate. No indication from software and no attempted restarts.

I've removed the --asm and the hashrate is now 60-70% of before. But within 30 seconds one of my GPUs randomly stopped mining and the hashrate indicator still says it is. Op. Now the other one has stopped, within 60 seconds of starting both.

Dammit. And I was all excited to let this run overnight after getting everything working.

Can anyone recommend another fork? I tried sgminer but the hashrate was under 10MH/s even after optimizing the intensity.

Edit: seems ok after installing every version of C distributable from 2005-2013 (2015 was already installed). 10 minutes testing so far. Will report back tomorrow maybe. Note that the previous issues had zero messages, popups or apparent crashes.

Further edit: one machine died overnight but that's because it was a fresh install - windows 10 default power settings standby after 30 minutes. The other machine was fine. So this seems to have solved my issue. Thanks for the optimized miner, OP.

Further, further edit: Added another machine and it had the same issue. Installing all the C++ distributable didn't help - quit after an hour. Only real difference is I had to install an older version of 10 (1607) due to circumstances. We'll see if upgrading to 1709 helps.

That all said, clearly some kind of failure-detection and recovery would be very much appreciated