DAMN! I'm really disappointed 
This miner was looking so promising. I have been using XMR-Stak for months now to mine (mostly) Loki using (mostly) cryptonight-heavy.
I am running Windows 8.1 X64 on a core i7 7700K, with 1 x gtx 1080 and 3 x gtx 1070s.
I have used Nvidia Profile inspector DmW to force P state 0 and I run all the cards at core +120 memory +500 (except the 1080 gets memory +400).
XMR-Stak gives me these hashrates: 1080 ~ 680 h/s 1070s 720-760 h/s
..so you can imagine my excitement when I tried CryptoDredge and immediately I see: 1080 ~ 800 h/s 1070s 880-900 h/s !!!
WOOT! I said to myself - I've found my new favourite miner.
BUT as always seems to be the case it was too good to be true.
I have a random system hard-reboot. Sometimes the miner runs for just a couple of minutes. Once it accepted about 100 shares and must have run for about 20 or 30 minutes. This was after my driver update but with my full original overclocks.
I have tried updating the Nvidia driver from a 397.xx version to a 398.xx version. I am avoiding going into the 4XX.xx drivers because I recently updated to 417.xx and saw a drop of 20% in XMR-Stak.
I have reverted back to P-state 2. I have lowered my overclocks a bit. Both of these tactics reduced the hash-rate advantage and actually seemed to make no difference to the reboot occurence.
Has anyone else seen anything similar? Any suggestions? XMR-Stak will run stable for days. If I had kind of stability and the hash-rates I see when it does run I would be a new convert and I would also go tell any Nvidia XMR-Stak users who would listen that they are missing out.
Unless this is a sudden coincidental hardware fault, it just seems to happen using CryptoDredge 0.9.3.
This miner was looking so promising. I have been using XMR-Stak for months now to mine (mostly) Loki using (mostly) cryptonight-heavy.
I am running Windows 8.1 X64 on a core i7 7700K, with 1 x gtx 1080 and 3 x gtx 1070s.
I have used Nvidia Profile inspector DmW to force P state 0 and I run all the cards at core +120 memory +500 (except the 1080 gets memory +400).
XMR-Stak gives me these hashrates: 1080 ~ 680 h/s 1070s 720-760 h/s
..so you can imagine my excitement when I tried CryptoDredge and immediately I see: 1080 ~ 800 h/s 1070s 880-900 h/s !!!
WOOT! I said to myself - I've found my new favourite miner.
BUT as always seems to be the case it was too good to be true.
I have a random system hard-reboot. Sometimes the miner runs for just a couple of minutes. Once it accepted about 100 shares and must have run for about 20 or 30 minutes. This was after my driver update but with my full original overclocks.
I have tried updating the Nvidia driver from a 397.xx version to a 398.xx version. I am avoiding going into the 4XX.xx drivers because I recently updated to 417.xx and saw a drop of 20% in XMR-Stak.
I have reverted back to P-state 2. I have lowered my overclocks a bit. Both of these tactics reduced the hash-rate advantage and actually seemed to make no difference to the reboot occurence.
Has anyone else seen anything similar? Any suggestions? XMR-Stak will run stable for days. If I had kind of stability and the hash-rates I see when it does run I would be a new convert and I would also go tell any Nvidia XMR-Stak users who would listen that they are missing out.
Unless this is a sudden coincidental hardware fault, it just seems to happen using CryptoDredge 0.9.3.
xmr-stak 2.5.1 vs CD 0.9.2 on cnv7 (3300 vs 3580)
https://imgur.com/a/B9HubsI
this is what optimized xmr-stak can do in cnv8 (3200). 1080 Ti, 1080, 1070 Ti and 1070.
https://i.imgur.com/8e4cVeD.jpg
OK so what I can see from your info is that you have core clocks lowered considerably and memory clocks as high as possible.
I might try that but it seems that no matter what i do, XMR-Stak is stable and keeps running but with a much lower hash rate than CryptoDredge - but when I run CryptoDredge, even with all GPUs at factory settings, my system reboots.
I was wondering, does the CD watchdog cause reboot or just stop the miner.
I was also wondering if I should just install Ubuntu and run CD in linux to see if that stops the goddamn rebooting.
OK so something occurred to me. I was originally using this machine for gaming so the display was coming from the 1080. I have now gone into bios, activated the onboard graphics and I'm using that for display, so now all the GPUs are just running cryptodredge and the 1st GPU isn't trying to mine and provide display simultaneously.
I hope this may solve my reboot problems but I will let everyone know. Rookie mistake maybe? I know that when I used to mine ETH I always had onboard video disabled and used the 1st GPU for display and mining but I don't remember the rules for Nvidia and custom versions of ccminer like this one.