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]

Also some people ditch the sata to 6 pin connection going into the riser and just use 6 pins off of their power supply if they have enough to spare.

On Tue, Dec 12, 2017 at 2:07 PM, Michael Ritten mritten@gmail.com wrote:

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 .

[minorminer59]

You don’t need to power the riser?

[mritten]

No they use a 4 pin directly off their power supply instead of the sata connector.

On Tue, Dec 12, 2017 at 2:34 PM, minorminer59 notifications@github.com wrote:

You don’t need to power the riser?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-351168480, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLtTu6zWj3jtZk9vFHZvyjWOLQjHcks5s_tU-gaJpZM4OEOmd .

[minorminer59]

Yes my risers are 6 pin to sata and my gpu is an 8 pin. So I can ditch the riser power?

[minorminer59]

Update- with 6 cards running at 100% power and no settings changes it seems to be running stable. However this isn’t great either 3 cards are doing 19 MHs and 3 are doing 15.

What are some next step trouble shooting? Also as a note windows seems to have a bit of trouble booting with 5 cards. If I swap pci slots or boot with 5 cards then reboot with the 6th it seems to work.

[ghost]

Hey guys,

so I had exactly the same issues and I think I fixed it. Since your comments helped me out a lot, I wanted to write what I did so far. Here is my system specs:

Asrock Z270 Pro4 PSU bequiet 1000W 6x Gigabyte GTX 1060 3GB Intel Core i5-7500 CPU Prozessor, 4x 3.40GHz Ballistix Sport LT 8GB DDR4-2400 CL16 rot SR

Before somebody is asking: I also use this setup for gaming ;)

I was using XMR_Miner with 36 blocks and 16 threads, which works usually good for gtx 1060 3gb.

What happened: I was setting up the system for the first time. I was using 4/5 PCIe and 2/2 M.2 interfaces for the 6 GTX 1060. It worked perfectly for one day. After I was forced to install Windows 10 1709, system was crashing all the time and screen frozen after about 2 minutes of mining. Then I rolled back to the last windows update, but somehow there still needs to be a driver issue that I cannot resolve anymore (even installed new drivers for cards), since windows now installs 1709 be default (even when creating a boot USB). This somehow results in a conflict of drivers, since using only one M2 interface works perfectly!!

How did I resolve the issue: I discovered that one of the cards was having about -5% hashrate compared to the other cards. So I decided to switch the cards around since there is one spot free on the mobo. I am using now 5/5 PCIes and 1/2 M2 interfaces. Fortunately (fingers crossed) nothing happened for about 24 hours, while having crashes in time frames of minutes before. The M2-PCIe adapter works fine since I switched them in order to verify this. At least to my knowledge, it has to do something with the Windows 10 1709 update and a resulting driver conflict for the M2 interfaces, which leads to lower performance of one of the M2s. Thanks for nothing Bill...

I can even overclock and undervolt now, without any issues.

Thanks to everybody here!!! Sluicifer is mining again :D

Greetz

[Lostinbios]

tried almost all in this topic ...i have my 1050 ti rig (5 x pgu) near a window and I experience that when i close the window my rig does the same... could it be heat ? i mean the risers ARE cheap china crap and when capasitors of the cheapest kind gets hot, they fail... Im testing out more heat issues today ! running succesfully for 14 hours now with the window open Copenhagen Denmark is 2 degrees celcius outside (35 farenheit) stay tuned :-) .......update I think i can rule out heat... it crashed with the window open.. I then skipped the x16 port and only used x1 PCIe ports and the whole computer now runs much smoother and no crash so far 20 januar 2018

[dragon9820]

Rig: MB: MSI Z270A SLI PLUS, CPU: i5 7500, 8GB ram, WIndows 10 PRo, HD: 64 GB SSD, PSU: 1050W Seasonic snow silent, GPU: 4 gtx1070& 2 gtx1070 Ti Powered Risers 3 cables to PSU Sata, Powered GPU's 6+2 pin to pcie on psu Display is ran by 1st gpu with vga to hdmi adapter.
BIOS settings: PCI 0 - Gen2, PCi 1- Gen2

After benchmarking I came down and the screen was off, I believe windows had froze, So held the button and restarted, still not screen. I had to go into BIOS and disable the 4g crypto setting then my screen would turn on. But with that setting disabled only 4 cards are able to use, 2 have code 12, stating not enough slots. Now with just 4 gpus with the 4g settings disabled I am able to start mining for about 10 minutes before either windows freezes or reboots. I have a fan pointing right at the rig and looking at the temps in Afterburner, they are between 50 and 65 degrees Cel. Just curious if you guys have any thoughts? Thanks!

[mritten]

Try running the display off the motherboard and try benchmarking first without overclocks. Then slightly begin to adjust the cards one at a time.

On Wed, Jan 24, 2018 at 3:26 PM, dragon9820 notifications@github.com wrote:

Rig: MB: MSI Z270A SLI PLUS, CPU: i5 7500, 8GB ram, WIndows 10 PRo, HD: 64 GB SSD, PSU: 1050W Seasonic snow silent, GPU: 4 gtx1070& 2 gtx1070 Ti Powered Risers 3 cables to PSU Sata, Powered GPU's 6+2 pin to pcie on psu Display is ran by 1st gpu with vga to hdmi adapter. BIOS settings: PCI 0 - Gen2, PCi 1- Gen2

After benchmarking I came down and the screen was off, I believe windows had froze, So held the button and restarted, still not screen. I had to go into BIOS and disable the 4g crypto setting then my screen would turn on. But with that setting disabled only 4 cards are able to use, 2 have code 12, stating not enough slots. Now with just 4 gpus with the 4g settings disabled I am able to start mining for about 10 minutes before either windows freezes or reboots. I have a fan pointing right at the rig and looking at the temps in Afterburner, they are between 50 and 65 degrees Cel. Just curious if you guys have any thoughts? Thanks!

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-360262448, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLpoKz16V7dn22E3w4sFQP8tTIYUaks5tN5H1gaJpZM4OEOmd .

[dragon9820]

For some reason I cannot get the display to come on plugged itno the integrated graphics card once the gpu's are plugged in. If I unplug all gpu's from the MB, then I can use the hdmi slot on MB. Does that make since?

[mritten]

You need to go into your BIOS settings, change the primary display settings to IGFX or onboard, whichever the settings are. Also make sure your BIOS is up to date.

On Wed, Jan 24, 2018 at 4:02 PM, dragon9820 notifications@github.com wrote:

For some reason I cannot get the display to come on plugged itno the integrated graphics card once the gpu's are plugged in. If I unplug all gpu's from the MB, then I can use the hdmi slot on MB. Does that make since?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-360272250, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLuRexYeo24ggUhEPY6AaYKgtLgDTks5tN5qBgaJpZM4OEOmd .

[dragon9820]

ok i'll double check that thanks

[killdog99x]

dragon9820 from my own experience 1070's can pull up to 200w each..your powersupply just may not be keeping up. causing the freeze

[dragon9820]

well after another long night, still no success I did change the display setting to IDG that definitely seem to help my display issue. I thought I had it at all working at one point after I disabled the cpu and just was running the 6 gpu, changed the pci settings to gen 1, nice hash ran for 20 minutes then I checked back and had the blue screen of death. Then i plugged them in one by one starting up and shutting down, when i got to the 4th slot, big 16 slot, display would not come on, no matter which gpu I plugged in. So I just had the 3 plugged in to see if I could get stable. Same issue and changing the pci settings back forth now the miner will go for a about 4-5minutes and windows just locks up.
So I would think the psu should have no issue power 3 gpu's, could it be my hard drive?? its only 64gb and I've read a lot saying to crank your virtual memory could it be too much heat, afterburner says around 60-65 degrees pci bus clock latency: is that very sensitive, i've changed it from 32,64,96
going through event viewer I did find an error nvlddmkm quite a few times

[banshee1971]

I build 1 RIG with MSI B250 PC MATE + Kingston ValueRAM 4GB (1x4GB) DDR4 2400MHz DIMM (KVR24N17S6/4) +
Intel Celeron G3900–2,8 GHz + EVGA SuperNOVA 55 G2 850 W Fully modular I have 3 x MSI RX 580 Gaming X 8Go and 1 MSI RX 580 ARMOR 8 GO OC + OCZ 120 Go SSD (old one i already got doing nothing at that time) Installed with Windows 10 Home (all windows Updage) + AMD Crypto Driver + Change all BIOS to optimized it ! Running fine for 2 weeks now at 630 W ! With EtherControl (to control remotly the profile, and see in action working).

Now i try to configure a RIG for a friends, using exact same component (except the PSU are EVGA 850GQ and Samsung 250 Go SSD + Windows 10 Pro. 3 x MSI RX 580 Gaminx X 8 Go.

System crashing all the time. I unplug the Riser of one card, seem crashing. But if i use 1 card at a time, very stable...

So i completly unplug 1 GPU (not only the riser), and start Claymore 10 (start.bat) instead of from Ethercontrol ... seem stable now for fiew hours.

Something tells me it's the power supply that seem not capable of handle all GPU...

Now next step, is to wait 24 hrs. If Stable, just plug the second card in the GPU connector only, and re-test it ! So valid it's power issue.

Seem that, from this thread, that all Windows freezing issue are power related most of the time.

[Alexisgarci]

Im having the same problem. I have 6 1080ti GPUS, z270 pro carbon, i7-7k, 32GB, 2 PSU's 1200W, windows 10 pro, all drives up to date and windows to. Windows freezes when i activate the 6 gpu. I can run 5 gpu's for days but when i turn the 6 onn on nicehash to mine, windows freezes after a while. I already changed risers, changed the order of the PCIE's, virtual memory is at 66GB at up 80GB, i tried GEN1 Gen2 in PEG01 and PEG 02. Anyone knows how to stop this freezes?

[mritten]

Are the cards overclocked at all?

On Thu, Feb 1, 2018 at 4:04 PM, Alexisgarci notifications@github.com wrote:

Im having the same problem. I have 6 1080ti GPUS, z270 pro carbon, i7-7k, 32GB, 2 PSU's 1200W, windows 10 pro, all drives up to date and windows to. Windows freezes when i activate the 6 gpu. I can run 5 gpu's for days but when i turn the 6 onn on nicehash to mine, windows freezes after a while. I already changed risers, changed the order of the PCIE's, virtual memory is at 66GB at up 80GB, i tried GEN1 Gen2 in PEG01 and PEG 02. Anyone knows how to stop this freezes?

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-362401136, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLgBW5xJjNrUQFi4qli6kb3cC9a_yks5tQibtgaJpZM4OEOmd .

[Alexisgarci]

Nop, i dont overclock, them, and ive tried underclock them.

[mritten]

How do you have nicehash configured? Is it mining the most profitable coins or do you have it set on a specific algorithm? I would try mining different algorithms and test each one for a few days to determine if it's a specific one causing windows to freeze. Another thought is to try and plug the risers in directly with a 6 pin connector from the PSU instead of the sata adapter to see if this makes a difference.

On Thu, Feb 1, 2018 at 4:08 PM, Alexisgarci notifications@github.com wrote:

Nop, i dont overclock, them, and ive tried underclock them.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-362402280, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLgKzD7A_qXT9Nag9kWNz7DFrrhK4ks5tQifagaJpZM4OEOmd .

[dragon9820]

yes this issue is driving me nuts, still trying to figure it out, my next step is to drop down to windows 1703 version and see if that helps, but I've tried alllll kinds of things between drivers and BIOS settings
getminer, I have noticed when it works well for a long period of time, its mining using hashimoto ive noticed it freezing more frequently with others, whether it be equihash or neoscrypt

[Alexisgarci]

Its connected directly from psu to the riser, no adapters.

[Petaflopminingco]

I THINK I HAVE THE SOLUTION. I was having this same problem on four different rigs, even one with just 4 gpus. MSI z270 gaming m5 MBs, 1080ti rig, 1060 6gb rig, 1050ti rig, and 1060 3gb rig.

So I redid everything from scratch, this time I installed etherminer first, started running it and then added all the other apps I use on my rigs.

As I was installing MSI afterburner the rig froze up. I restarted it, uninstalled afterburner, and then started etherminer again. So far its going well for the last 15 min. used to freeze after 3 min.

[Alexisgarci]

but you dont use nicehash?

[dragon9820]

With uninstalling afterburner. Won’t you lose your overclocks? Or not

[Alexisgarci]

yes there is another programs to use. Ill give it a try tomorrow. Every time i set all 6 to mine it gives me the Blue error of video tdr failure nvlddmkm.sys. I think i already saw every vid on youtube about it and none works :/

[Petaflopminingco]

UPDATE. Its not afterburner, afterburner just starts using more CPU, and if youre not configuring the batch file exactly right for your hardware you will use too much CPU.

Run task manager while you run your miner and see if the CPU usage stays high. If it does, the problem is that your cpu is being overuseed. Search the cpu issue with your hardware to find a solution. For me, using nvidia, I had to switch the -G in etherminer to -U . Working great now, CPU staying below 20.

[banshee1971]

My cpu is a celeron, and dont go more than 10%. When windows freeze (usualy after1-2hrs) there is no clé on what happen (3 GPU app open to monitor GPU, and task monitor in performance monitor mode). The freeze screen show me normal behavior.

I got no bluescreen and no error on the event viewer.

[banshee1971]

The only différence between those 2 rig : windows pro and home edition (home edition stable with 4 GPU)

I use msi rx 580 game x 8g

[banshee1971]

Im using claymore Dual (used in ether mode only)

[Petaflopminingco]

Well this seems to be working for me, of course it hasnt been running that long though. My cpu was spiking, but now its staying low. Again, check your batch file settings, make sure it matches your hardware and OS. The little letters make all the difference.

[Alexisgarci]

where do we check the batch file? i have an i7-7k and i mine with it, so its not the cpu in my case. What windows do you guys use?

[Petaflopminingco]

I use i3s across the board. Windows 10. If your CPU is at 95% or more at any point, it is likely the CPU. otherwise it could be something else.

[Alexisgarci]

Windows home or pro?

[Petaflopminingco]

The batch file is your start file for the app. Its in the folder with the app. Using pro.

[dragon9820]

I’m using nice hash I’m really wondering if it’s a windows issue. When I back it down to 1703. I’ll let you know if it still freezes every 1-2 hrs like before

[Petaflopminingco]

When I was using nicehash it kept crashing, so I disabled dagger-hashimoto and that seemed to work. I see a lot of people talking about one of the NH currencies causing the crash.

[dragon9820]

Interesting. Because that’s what algorithm Im always using.

[minorminer59]

I had all of these crashing issues too. I ended up swapping all of my risers for the latest ubit ones. I blew out all of my drivers and reinstalled. I have been running smooth for almost two months now. Zero system interupts.

Open up your activity monitor. If you see fluctuations in your system interupts more than likely bad risers.

[dragon9820]

Well just an update here. I’m using the latest nice hash miner version and also use legacy Windows usually freezes within 1-3 hrs. I went ahead and turned windows updates back on and and installed update KB4058258. Been up and running for 11 hours straight!!!

[Alexisgarci]

I cant turn the 6 GPU on in nicehash, if i turn it on ill get the video tdr failure nvlddmkm.sys blue error after a while. Im mining with 5 gpus only :/ sucks

[dragon9820]

yea, definitely! does your board have the 4g crypto option in BIOS?

[Alexisgarci]

yes, i already turned it on and off. You guys have Peg0 and peg1 or only peg0?

[banshee1971]

I have 4g crypto option in BIOS, and on my working Miner, i have only the PEG option enable. Working stable since i finish adding the latest GPU (was stable all the time anyway) here the XLS sheet i create to document my BIOS config of this MSI motherboard !

The 4 G option only appear since the latest version of the board. I upgrade the BIOS of both RIG, and this as no effect on both ... mine still stable, the other still crashing !

On the Crashing RIG ; I change the option between PEG and the other option (to have the onboard card on), in case it improve things, play arroung with the 4 G option (enable/disable),...

The board on both RIG are MSI B250 PC MATE Both has the same brand /model of CPU and memory. But now my friends is ordering a 1000 W PSU just in case his 850 W are faulty, and he just receive his 4th GPU (so the RIG will have 4 instead of 3 GPU).

I heard a comment regarding ethermine working better with pair of GPU... is it possible that it's the case with Claymore ? (because my issue are with 3 GPU... 2 Works like a charm)

The other, since this faulty RIG has Windows PRO ... and i know that there is more option activate in PRO than HOME... is it possible also that it's there that reside the issue ?

I plan to perform a test this weekend... install on a spare drive Home edition just to valid this.

I'M 100% sure about 1 thing, the GPU and GPU riser works fine because i made them work strait 48 hrs without issue.

[Alexisgarci]

We have diferent issues. Mine don’t freeze I’m just getting the blue error on video_tdr. I’ll try your setting and I’ll see if I get a crash. When I get home I’ll do this and I’ll share with you my settings. This shit is getting me crazy. I think it’s my mobo I have a msi z270 pro carbon. I’ll have to try another mobo to see. I’m considering switch to Linux tho.

[mritten]

Nvidia cards should use a PCI link speed of GEN 2, AMD Cards use a PCI link speed of GEN 1.

On Fri, Feb 2, 2018 at 1:50 PM, banshee1971 notifications@github.com wrote:

I have 4g crypto option in BIOS, and on my working Miner, i have only the PEG option enable. Working stable since i finish adding the latest GPU (was stable all the time anyway) here the XLS sheet i create to document my BIOS config of this MSI motherboard ! [image: 2018-02-02_13-20-35] https://user-images.githubusercontent.com/7092486/35748399-eab3c124-081b-11e8-9925-d9c67b010e2c.png

The 4 G option only appear since the latest version of the board. I upgrade the BIOS of both RIG, and this as no effect on both ... mine still stable, the other still crashing !

On the Crashing RIG ; I change the option between PEG and the other option (to have the onboard card on), in case it improve things, play arroung with the 4 G option (enable/disable),...

The board on both RIG are MSI B250 PC MATE Both has the same brand /model of CPU and memory. But now my friends is ordering a 1000 W PSU just in case his 850 W are faulty, and he just receive his 4th GPU (so the RIG will have 4 instead of 3 GPU).

I heard a comment regarding ethermine working better with pair of GPU... is it possible that it's the case with Claymore ? (because my issue are with 3 GPU... 2 Works like a charm)

The other, since this faulty RIG has Windows PRO ... and i know that there is more option activate in PRO than HOME... is it possible also that it's there that reside the issue ?

I plan to perform a test this weekend... install on a spare drive Home edition just to valid this.

I'M 100% sure about 1 thing, the GPU and GPU riser works fine because i made them work strait 48 hrs without issue.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-362665993, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLrLoa-PGdGdTJJKfOa9buEICJ5k3ks5tQ1PHgaJpZM4OEOmd .

[Alexisgarci]

Ive tried both. I hve peg0 and peg1 ive tried gen2 in both gen1 gen3 and cant still add the 6 gpu

[mritten]

Not all motherboards fully support 6 GPU's even if they have enough PCI slots. It's an issue with a few motherboards that may have 6 PCI slots, only 5 GPU's will work. That's why the Asus Z270 boards are so popular.

On Fri, Feb 2, 2018 at 2:09 PM, Alexisgarci notifications@github.com wrote:

Ive tried both. I hve peg0 and peg1 ive tried gen2 in both gen1 gen3 and cant still add the 6 gpu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-362677240, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLtqr4dSvMk5MiJ9kRLvm3-fs2J3hks5tQ11xgaJpZM4OEOmd .

[Alexisgarci]

i have a msi z270 pro carbon, ive seen already people using it with 6 gpu but seems i cant do it :/

[mritten]

Google your motherboard and you will find out that it's a common issue with mining with 6 GPU's. It's either hit or miss with MSI. Have you tried using a M.2 adapter to see if that makes any difference?

On Fri, Feb 2, 2018 at 2:49 PM, Alexisgarci notifications@github.com wrote:

i have a msi z270 pro carbon, ive seen already people using it with gpu but seems i cant do it :/

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nicehash/NiceHashMiner/issues/1085#issuecomment-362688038, or mute the thread https://github.com/notifications/unsubscribe-auth/AcTrLtghxnFW988SA3TUUnSYbITqNwoeks5tQ2bCgaJpZM4OEOmd .

[Alexisgarci]

havent tried the m.2 adapter yet. Ye i think is that, or we are lucky to get one mobo who works or one who cant handle it :/