Mining rig is unstable and cannot understand why, help me please!

[murgorx]

Hey guys! For the last couple of days I have been having a lot of issues with my mining rig.
I have 5x RX580 sapphire nitro+s, which are usually running really solid, but after 5:30 hours of mining I receive the  following issue: https://imgur.com/a/InCeI .
As you can see my #4 GPU is down and after the reset I get those messages.
Yesterday I had the same problem with #2 GPU if I am not mistaken. I lowered the cclock to 1150 and the mclock to 2100, since yesterday's issue, but I am still getting the same problem. It is constantly happening and I don't know why, but both times after around 5hours and 30minutes of mining.
Has anyone seen anything like it? Could you guys please advise me how to proceed? Because of that I have around 7-8 hours downtime everyday and it getting really annoying...
I cant leave it overnight or when I'm work.

[1715101984]

1715101984

[1715101984]

1715101984

[1715101984]

1715101984

[mshordja]

Hey guys! For the last couple of days I have been having a lot of issues with my mining rig.
I have 5x RX580 sapphire nitro+s, which are usually running really solid, but after 5:30 hours of mining I receive the  following issue: https://imgur.com/a/InCeI .
As you can see my #4 GPU is down and after the reset I get those messages.
Yesterday I had the same problem with #2 GPU if I am not mistaken. I lowered the cclock to 1150 and the mclock to 2100, since yesterday's issue, but I am still getting the same problem. It is constantly happening and I don't know why, but both times after around 5hours and 30minutes of mining.
Has anyone seen anything like it? Could you guys please advise me how to proceed? Because of that I have around 7-8 hours downtime everyday and it getting really annoying...
I cant leave it overnight or when I'm work.

have you change the riser, or even the PCI port. I had the same problem time ago and resolve by only changing the PCI port to another, or switching with another card. Give it a try, and of course check the PSU

[murgorx]

The problem is that it happens on different cards , not only on one card mate and I do not have other pci-e ports left, since from now i'm starting with 6 gpus. May it be a problem in the BIOS mods or overclocking/undervolting? Im running on 1150/2150, which imo should not be causing any issues.
#I'll try with a different riser to see if thats the case. Thanks for the answer

[asukahan]

test them with no OC before saying no problem to those OC profile

[Set Ready Go]

Like others have said , try set everything stock settings , if stable 48 hours , then you know its about the uv/oc

What risers are you using?
Because like you said its diffrent gpus,  i was thinking maybe it is power distribution problem.
So then to the question: What Psu/psus are you using?   number of risers you power per strand molex/sata/pcie

[murgorx]

PSU - EVGA SuperNOVA G2 1300 x2 PSU PC POWER SUPPLY (G21X, 12 V, KM, Eco Mode, 1300 Watt
Cards - Sapphire Radeon RX 580 Nitro Special Edition 8GB GDDR5 2 X DP 2 x HDMI/DVI-D Graphics Card – Blue
Risers - https://www.amazon.de/Ptsaying-Powered-Adapterkarte-Verl%C3%A4ngerungskabel-Netzkabel/dp/B073TDV2ZM/ref=pd_sim_147_5?_encoding=UTF8&refRID=EB6WVS5Q7QV2XFSJSZ0M&th=1 SATA powered.

I haven't had this issue for the first two weeks while mining with this equipment.
The whole system is drawing 700w when the PSU is 1300w, which imo should not be the case.

[Set Ready Go]

PSU - EVGA SuperNOVA G2 1300 x2 PSU PC POWER SUPPLY (G21X, 12 V, KM, Eco Mode, 1300 Watt
Cards - Sapphire Radeon RX 580 Nitro Special Edition 8GB GDDR5 2 X DP 2 x HDMI/DVI-D Graphics Card – Blue
Risers - https://www.amazon.de/Ptsaying-Powered-Adapterkarte-Verl%C3%A4ngerungskabel-Netzkabel/dp/B073TDV2ZM/ref=pd_sim_147_5?_encoding=UTF8&refRID=EB6WVS5Q7QV2XFSJSZ0M&th=1 SATA powered.

I haven't had this issue for the first two weeks while mining with this equipment.
The whole system is drawing 700w when the PSU is 1300w, which imo should not be the case.

Sounds strange indeed.

Have you set virtual memory to 16 gb?

and the rest of the settings for recommended multi gpu rig?


Also you can try and DDU the rig and install latest Adrenalin drivers. 
Easy for u to set the OC and undervolting in wattman there and save profiles.

[murgorx]

I am with the blockchain drivers from 17.12.1 drivers. I just checked the Radeon settings and saw there's an update 18.1.1.
Should I be using the adrenaline drivers instead of those?

[Amstellodamois]

What's your voltage profile?
I'd try with 1200 GPU clock to see if it's more stable.

Also, when your miner hangs and try restarting, have you had it restart the whole system? Not ideal but would hopefully limit the downtime.

[murgorx]

When I receive this error and I try to close the miner it wouldn't do it and I have to make a hard reset of the rig.
Today I was at work when I made a TeamViewer session to check the miner and saw it - tried to click on the Windows button, but no response.
Went home after 5 hours - it was still at the same damn position and the only way to restart it was to turn off the power button on the PSU. After a couple of seconds I turned it back on and the rig started up again.


As for the Windows Settings - I have followed everything, that's in this article - http://1stminingrig.com/best-windows-setup-configuration-tweaks-for-mining/.

[leonix007]

I would suggest to try these for isolation, Trim down your GPU to 3 or 4 only if your board supports 4 PCI-e x16 then use this without any risers. use only 1 PSU, if your rig runs solid after this actions, then, probably the problem is in your Riser or PSU.

[Amstellodamois]

For a remote hard-reset, buy one of these: https://www.amazon.com/s/ref=nb_sb_noss?url=search-alias%3Daps&field-keywords=wemo+insight

And try new settings as advised.

[murgorx]

Okay, those gadgets seem really nice! Thanks.
As far as the drivers - Should I be with the current ones - blockchain 17.12.1 or should I upgrade to 18.1.1 or should I install the Adrenalin drivers.

[Amstellodamois]

I'm using the latest Adrenalin drivers (in compute mode) and have a solid 31+ MH on every card.

Please pay attention to my post #9, it might help you.

[murgorx]

I was running on 1200/2200 the first two weeks and it was stable AF, just afterwards I started getting some crashes or tried to do other "tweaks" and that's why I was trying with 1150/2150.
Now Im running nicehash just to check what would I get from them and it seems that some cards are performing waaay better than others?!
I am thinking of buying a BIOS mod for my sapphire nitro+ micron cards from the guy Mattthev and to try modding them all with it. Maybe I've done something wrong with the BIOS mods, although I have only copy-pasted a single line from 1750 to 2200 But just for my own sake and peace.

[murgorx]

As far as Risers, could you please tell me if my are okay? Or just give me your opinion on which should I buy? Would prefer SATA ones, just like mine.

[Amstellodamois]

Again (and for the last time): post #9

[murgorx]

Again (and for the last time): post #9

Amstellodamois, I am sorry for the silly question, but should I create a screenshot from HWINFO to show you? Is that what you want to know? I'm not able to understand you fully, sorry :S

[Amstellodamois]

Did you modify the voltages of your cards? If so, they can be unstable.
Try 1200 MHz for the GPU, you'll gain hashrate and stability (and consume more power).
Check Claymore's tutorial: you can run a batch if the miner fail, have that batch restart your rig.

[murgorx]

No, haven't done anything like it, because I've got no knowledge and was not really sure what to do, so left it as it was.
I just installed the 18.1.1 drivers and have gone through patching 2 cards - the first one is 31.5mh/s and the second one is 30.3-5mh/s.
Will go through all of the 5 now and see what's up.
Changing also to compute mode, but it is not really doing much as in mh/s improvement.