6 GPU miner freezing windows. Runs fine for a while then stops?

[hungryhungarian]

Im running 6 GTX 1080 ti ZOTAC AMP Extremes. Four GPU's run like a charm with no issues. Five I havent tested extensively but when I did it seemed fine. I will do more test soon and update. However with all six, stability issues or something of the sort come up. It will run Nicehash for a good 30-45 minutes and do really well, and then all of the sudden either freeze windows or the prompt will repeat "unknown error" over and over.

Ive done everything I can think of and stuff that Ive found online. ANY help would be greatly appreciated. Ill even send you eth or something if you help me resolve this frustrating problem.

[mritten]

Having the same issue here, have 4 GTX 1060's and 1 GTX 1070, windows 10 is up to date and using the most recent nvidia drivers. Nicehash will run daggerhasimoto for a few hours and then windows will restart or will lock up. I'm going to re-install windows 10 and start over from scratch to see if that fixes the issue for me. All benchmarks work fine. My other system with just 2 1060 Zotac's mine daggerhashimoto all day long without any issues.

[hungryhungarian]

I actually made a small breakthrough today. I have been testing only using 5 GPU and it was doing the exact same thing as when I used 6. I unplugged and put the PCIe risers in a different order on the motherboard, and now the 5 GPU setup has been running flawlessly. No idea if this is a fluke or what but I'm hoping when I plug the 6th in it'll run.

[mritten]

I wonder if your issue may be related to a riser being bad? Also what is your bios PCIE Gen set at? I changed mine to Gen 2 and I still had the issue of system restarting or freezing.

[hungryhungarian]

I used the same risers, but in a different order so it shouldn't be that. I have mine on gen 1. Also I should not that I put the HDMI out put on another card that switched pcie spots with the original. If I'm not mistaken this actually solved some power fluctuation issues with the first card.

[mritten]

What size PS are you using?

[hungryhungarian]

Using 2 psu just to be safe, one 1600w powering the motherboard and 4 GPU's. And one 850w powering the other 2 psu's. It's overkill but just barely.

[mritten]

Ok just wondering if it was somehow under powered and that was the issue.

[mritten]

A fix that I found to work is to increase the windows 10 virtual memory system size. It defaults to a low amount. If you increase it to an 8gb size of ram, I set it to initial: 8,192 and maximum: 12,288. Nicehash utilizes the virtual memory and can cause issues from what i've experienced if not updated.

[anonymoous]

I am having the same problem. It runs for a few hours then stops. I also have 6 1080's

My virtual memory is set to like 50gb so that's not an issue.

My PSU is a 1600w

[anonymoous]

Could this be a heat issue? Are you guys overclocking?

[hungryhungarian]

It definitely isn't heat for me, mine are running between 40° and 60° never over

[hungryhungarian]

I increased my virtual memory as well before posting this so it's not that, at least for me

[mritten]

I switched over to Claymore miner which seems to be more stable so far especially with my AMD and Nvidia card rig. I will know more after I have it run for a day or so, but so far the hashrate is staying stable where in Nicehash the ETH miner hashrate fluctuates and the miner would restart periodically.

[anonymoous]

Yeah i switched also because this Nicehash miner is just un-usable for me at the moment. Using ethminer

[anonymoous]

Do the devs read this? Would be nice to get a comment from them on this issue

[theLosers106]

Two things that I can suggest you should try is that split the 6 GPUs to 2 different NiceHashMiner instance (GPU1,2,3 NHM Instance 1 & GPU4,5,6 NHM Instance 2). Another thing you could try is that I had once had an issue running it for more than 1 hour and it was because of the RAM. Hopefully this helps.

[myrig]

Hi there i had the same problem with gtx 1070 any time that you put more display cards it require mor memory to send to processor. So i had 5 gtx 1070 than anytime that has big transactions to confirm your rig it crashes the system. So i solved by puting 4 more gb on motherboard and everything works fine hope this will solve your problems. Simply if you have 4gb increase it.

[hungryhungarian]

Update guys, I had ordered 2 8gb ram sticks for my personal PC and decided to try those out on the rig. It's been running for 3 days straight now. Still VERY glitchy, so once it started running I just straight unplugged any mouse/keyboard input and let it do it's thing lol. I check that it's still running on the nicehash website with MyMiner.

[p1r473]

So you just needed more RAM?

[hungryhungarian]

That was one of the problems, man to be honest I can't list the amount of things I tried. It just started working all of the sudden but yes try adding RAM

[mritten]

Just to confirm you already increased your virtual memory size which did not work for you?

[p1r473]

I am having serious stability issues. I am getting a lot of freezes (have to reset the power) and some TDR BSODs (I already increased TdrDelay to 60sec) and I am on the latest drivers I am hoping someone can help me. I just built a 6x 1080 TI rig (aorus gigabyte) and I am not using their software. The return window on my GPUs is closing fast, and I need to get this stable ASAP!

Yesterday I mined for several hours stable, so I don't think its a hardware issue.

Why do I keep freezing and BSODing??

I just ordered 2x 8gb ram (was using 2x4gb ram) as per your recommendation. What else do I try? Desperate to get stable! Was a huge investment for me, and it crashes every hour!!

Any more potential fixes I should try?

[hungryhungarian]

Yes I had already increased Virtual memory, no help at least for me.

I never got a BSOD so I can't help you there. Keep googling and try everything you find. Eventually you'll find the problem. I'll go through my bios when I'm home and share all my settings.

[p1r473]

Its mostly freezing, but some TDR BSOD too

[mritten]

Please review the previous comments on here.
What Power Supply do you have?
Fresh install of windows? Is windows up-to-date? Does your motherboard have a molex input near the PCIE slots? Sometimes a motherboard requires this to be plugged in when utilizing multiple cards. BIOS is up-to-date? Try PCIE Gen 1, Gen 2 Settings Latest Nvidia drivers only, do not install VR and other misc drivers. Disable motherboard audio as well as Nvidia sound. Virtual Memory size increased. Update windows power settings and ensure PCI is set on maximum power and disable other power save settings (Hybrid, Hard drive shut down) Mine on stock settings at first to ensure system is stable before beginning to overclock.

[p1r473]

Up to date fresh install of Windows. 2x 1250w psu. I did not plux the molex for the motherboard as I am using powered risers Bios is current. PCIE gen 1 and 2 tested Nvidia drivers only installed, no other software Virtual memory increased a lot Power settings max Overclock is only +0 Nvidia hdmi sound output disabled

Will try to disable mobo audio.... Only thing I havent tried yet

[hungryhungarian]

This may sound weird or pointless but try rearranging the order that the risers are plugged in. For some reason one of my GPU's really didn't like being the vid output and had large power fluctuations. Once I switched to another as output both of them had stable power usage.

[p1r473]

Will try that too thanks!

[mritten]

That is a good thought, or just try plugging in 1 card first and getting it stable, then add another card and so on until eventually you have all cards running. If it becomes unstable at any point when adding a card then you will be able to figure out if it's an issue with the riser on that card or something else.

[p1r473]

I also heard running 2 instances of NiceHash may work (3 GPU per instance) Will try this all....

[p1r473]

Also will try to disable non-profitable algorithms

[myrig]

Hi there. I am having many trubles with stability and now trying those options and lets hope if this solve the problem by the way i will keep updating my experience with you guys :) I have GTX 1070 1x6 pins & 1x8pins power consumption connectors

1. No more than 4 gpus per rig.
2. Incerease virtual memory to mycomputer/properties/advanced/virtual memory to atleast min & max 32000 mb.
3. Keep the PSU temperature low. most of the troubles is cousing the power feed.

Hope this soleve your and my problems.

[mritten]

Make sure you only are testing stability with 1 algorithm. Another way to test stability is to just download the latest Claymore 9.6 and set it to mine on the nicehash pool if that's what you prefer. This way you can monitor temps and see if the crashing is because of the nicehash software.

[myrig]

Guys i found a solution resatarting the rig atleast every 6 hours before the system crashes also new version of nicehash is better than old one. Find wich algorythm is causing the crash and disable them. I left only daggerhashimoto. Also it pays better than others.

[mritten]

Have you tried another miner out? Like claymore to see if it still crashes.

[smartmanjp]

There any fix for it yet i think everyone have the some problem i 4 Rigs 32 Cards all the 4 crash to many time every day i try everything i build computer all my life I know a lot of computers windows and Linux but i never see an systems crash so many time and i think there no fix for that if anyone have any way to to fix it please let me know

[pstealth]

Possible solution....I have the same issu with 6 Zotac 1080ti amp extreme...I found that one of my card on MSI AFTERBURNER was set to power limit 120. This card was causing the crash...All others card was at power limit 100....So check your power limit. The Zotac are already push at max stock. When I have install the last card, I have forgot to CUT the link between power limit and temp in msi afterburner for that last card only...So when I have push all cards to temps max 90 degres, the las card raise thepower limit to 120 by default.

[mritten]

Well for now Nicehash is down so you should be using a separate miner anyways.

On Thu, Dec 7, 2017 at 4:08 AM, pstealth notifications@github.com wrote:

Possible solution....I have the same issu with 6 Zotac 1080ti amp extreme...I found that one of my card on MSI AFTERBURNER was set to power limit 120. This card was causing the crash...All others card was at power limit 100....So check your power limit. The Zotac are already push at max stock. When I have install the last card, I have forgot to CUT the link between power limit and temp in msi afterburner for that last card only...So when I have push all cards to temps max 90 degres, the las card raise thepower limit to 120 by default.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-349906407, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLn1imwUPp2pboiJIsToCcIABuiFCks5s96rwgaJpZM4OEOmd .

[pstealth]

I use claymore now...mining Zcash and money send every hours in my trezor wallet, learn my lesson. (I loose 1 day of mining with Nicehash, 200$) I'm lucky..I have made a withdraw of 800$ one day before de Nicehash Hack.

[minorminer59]

I am new to mining and having all of these issues. Please advise if I need to create a new thread.

I am running 6 Evga gtx 1060 6gb SSC ( cannot find a stable OC settings.) 3 cards are samsung and 3 are Hynix. Two PSUs 850w and 1000w.

I was able to get 5 cards semi stable running 65w + +50core +500mem after about two days it would freeze the whole pc. Have to hard reboot.

I added a 6th card and cannot run for more than 5 minutes without it freezing.

I removed 6th card and still freezes with only 5.

I am using leboo risers. Is it possible there are issues with dual psus or maybe bad risers?

Where do I begin to trouble shoot this mess?

[mritten]

How many risers do you have plugged in to the same strand from the power supply? I found that anymore than 3 risers plugged into the same sata power cable from the power supply can create issues. To help narrow your issues remove the overlock and see how long the system will run before locking up. What OS are you using?

On Tue, Dec 12, 2017 at 1:09 PM, minorminer59 notifications@github.com wrote:

I am new to mining and having all of these issues. Please advise if I need to create a new thread.

I am running 6 Evga gtx 1060 6gb SSC ( cannot find a stable OC settings.) 3 cards are samsung and 3 are Hynix. Two PSUs 850w and 1000w.

I was able to get 5 cards semi stable running 65w + +50core +500mem after about two days it would freeze the whole pc. Have to hard reboot.

I added a 6th card and cannot run for more than 5 minutes without it freezing.

I removed 6th card and still freezes with only 5.

I am using leboo risers. Is it possible there are issues with dual psus or maybe bad risers?

Where do I begin to trouble shoot this mess?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-351136172, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLq1oyMTRuyN2x9mqN9l1D1mavOq1ks5s_sFVgaJpZM4OEOmd .

[minorminer59]

I have 4 risers on the 850w and 2 on the 1000w (the 1000 is new.)

I’m running Windows 10 Pro.

[minorminer59]

Also main board is MSI Z270 Gaming Plus with i7 -7700k 3600 - 32 GB ram. not sure if info was needed

[myrig]

Hi there i saw many threads about this problem. First increase virtual power. This will solve freezing problem. Seccond make a bat file to force your pc restart every 2 hours and put it on startup folder. Do the same by creating a shortcut of nicehash/claymore/cgminer or what softwer do you use. This is to start apps at startup of pc so you dont have to take care always turning them on.

Best oc settings for gtx1070 tem 70c 145 and the last 500-600 at afterburn depand on script that you load.

Hope this helps you guys

[myrig]

Hi there i saw many threads about this problem. First increase virtual power. This will solve freezing problem. Seccond make a bat file to force your pc restart every 2 hours and put it on startup folder. Do the same by creating a shortcut of nicehash/claymore/cgminer or what softwer do you use. This is to start apps at startup of pc so you dont have to take care always turning them on.

Best oc settings for gtx1070 tem 70c 145 and the last 500-600 at afterburn depand on script that you load.

Hope this helps you guys

[minorminer59]

My Virtual memory is set to 16gb I’m running claymore eth miner. The settings for 1070 work for 1060 6gB SSC? Why would I need to restart every two hours? I mean is this really necessary?

[myrig]

Depand to miner. Restart gives the psu and vga a little time to cool for a while. Most of times the power connectors get burned and sometimes affect the gpu. I have 6 months runing 24/7 no problem yet

[mritten]

Distribute the risers so that 3 risers and video cards are running off the 850w and 3 riser and video cards are running off of the 1000w. Try that out and make sure you have the latest driver and windows updates.

On Tue, Dec 12, 2017 at 1:31 PM, minorminer59 notifications@github.com wrote:

I have 4 risers on the 850w and 2 on the 1000w (the 1000 is new.)

I’m running Windows 10 Pro.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-351143957, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLn3-PO96wjnD105JS2vunEY6E0q4ks5s_sZpgaJpZM4OEOmd .

[minorminer59]

Ok thank you for the tips. If three risers per psu are recommended, how do people build 8 Gpu rigs? Just curious not doubting. Also, would a bad riser have anything to do with this? Or would a bad riser not work at all?

[mritten]

A bad riser could do it as well. Usually I will run 2 or more sata cables to connect 3 risers on a single cable each, but since you have 2 Power Supplies it's worth distributing the load to rule out any other issues.

On Tue, Dec 12, 2017 at 2:05 PM, minorminer59 notifications@github.com wrote:

Ok thank you for the tips. If three risers per psu are recommended, how do people build 8 Gpu rigs? Just curious not doubting. Also, would a bad riser have anything to do with this? Or would a bad riser not work at all?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-351156826, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLmE1EmUL6CWDiFoT5nftlryiW6DCks5s_s5rgaJpZM4OEOmd .