Hello,
I had the problem that system freezes at random time intervals while mining when using 10 Nvidia cards on H110 motherboard running on Win10. I was aware of the limit on Win10: 8 nvidia and 5 amd, or other way round so I got information that windows solved that problem in Fall creators update (1709). Unfortunately, my system was running two days and after that it started to freeze again on random time intervals when mining: approximately 2-3 hours. I tested both EWBF and DSTM miner. Has anyone reported that this win update can fix the problem?
This is a description my system conf:
Rig:
10x Zotac 1070 AMP Extreme
Asrock H110 Pro BTC
Intel i3
8GB RAM
500 HDD
2x corsair 1000W + Kolink 750W : 2750W
OC:
Power Limit: 65
GPU Core: 70
Memory Core: 270
Target Temp: 70
OS:
Windows 10 1709, Fall creators Update
EWBF: 4600 sol/s Power usage: 160W/card
DSTM: 4750 sol/s
I think that I have finally found what was causing the problem. It seems that the last PCIe is not working properly. When I insert extender with the card in it, Win freezes one minute after turning on. Still, have the issue with stability. It works stable only for approximately 15 hours on average. Maybe it's PSU.
Did you try a different raiser? It worked for me twice to fix an unstable rig.
FYI Almost all of the "bad raiser" issues come down to the bad connection of 12v within molex to raiser connector.
RE stability issues do try to disconnect and reconnect power to raisers, also various tasks that windows may start automatically may cause issues. (disable the bunch of Nvidia tasks in Windows Task Scheduler for starters)
What version of riser would be the best to use for stable mining?
I just bought 6 riser which is version 007 and the board is color red do you think it is safe if i use the direct cable from hxi 1200i corsair for sata power cable directly to riser? or do i still need to molex to sata cable?
After one month of struggling, I think I have finally solved the puzzle. It wasn't the non-working slot on the motherboard as I've previously thought. The solution is as follows:
I bought a new motherboard. I tested one by one GPU with risers. My first problem was random Win freeze. I found out that 2 of 10 risers have been causing that problem. Namely, with only one card on a new motherboard( with unhealthy riser) it turned out that windows also freezes. I removed those risers and the rig started to work normally.... Except for one additional problem:
Zm miner (dstm) or ewbf miner crash. ( longer time intervals on low clocking settings). I performed the same experiment now, tested one by one GPU, but this time I pushed overclocking almost to the limit. It turned out that I had to throw away two more risers.
To summarise: I noticed that two types of riser problems existed on my rig:
1. Bad Riser was causing windows to freeze, or restart with blue screen randomly.
2. Bad Riser was causing Miner to become unstable. ( crashed on longer time intervals).
One interesting thing that I also noticed was that one GPU from 10 that I had( ZOTAC AMP extreme) was unable to run on high overclocking settings( and all other cards were able). I noticed that when I tested the riser on that card. With riser, miner was crashing. I removed the riser, plugged card directly into PCIe 16x slot, and surprisingly the same happened without the riser.( that have never happened with other 9 GPUs) That made me conclude that I have got one card that definitely has worse SMOS integration done. ( Silicon lottery)
So I want to emphasize: when you are overclocking your rig, be aware that only one card can limit your max clocking. Test each card separately. Now I am using Acrefawn's watchdog:
https://bitcointalksearch.org/topic/cmd-farmwatchbot-ewbf-claymore-bminer-dstm-cc-eth-castxmr-phoenix-2071108 It helped me because still, I have some program crashes after few days of running. The batch script automaticly runs a program after a crash. It has also some more advanced features.
