Mining rig resets after 10 minutes

[labsy]

Hi guyz!

I have a new mining rig (MSI Z270-A, 8x KFA2 1070Ti, 2x 850W PSU) and have problems running all 8 cards.
When all 8 GPUs are running, rig simply after some 10-20 minutes shuts-off or just stops mining, depending on OS I use.

Settings:
- Above 4G ON, PCIe Gen1 or 2
- No overclocking

Tried:
- With windows 10, SMOS, HiveOS linux...same story
- Exchanged risers...same story
- Changed CPU cooler and thermal paste ....same story
- Changed RAM ....same story
- Underclocking ....same story

Symptoms:
- all 8 GPUs .... stuts-off after 10-20 minutes or simply stops mining
- only 4 GPUs on one PSU    ....works OK
- only 4 GPUs on other PSU .... works OK
- only GPUs 1, 3, 5 and 7 on both PSUs ....works OK
- only GPUs 2, 4, 6 and 8 on both PSUs ....works OK

Any idea?

[Turkish88]

Do you use overclock ? if yes try decrease clocks.
Run 7 gpu, swith of one by one, to detect problem card

[Geraldo]

From symptoms that you describe:

- only 4 GPUs on one PSU    ....works OK
- only 4 GPUs on other PSU .... works OK
- only GPUs 1, 3, 5 and 7 on both PSUs ....works OK
- only GPUs 2, 4, 6 and 8 on both PSUs ....works OK

Seems PS enough to handle your cards. Just for reference.

Silly question from me.
Are you using power switch or reset switch or both on your Mobo?

[somethingiswrong]

Are you powering more than two risers from one lane of your PSU?
You should not connect more than two risers to one lane.

[kunsh]

Check power supply or even better, you needto undervoltage your GPU's, there is detailed explanation on youtube how to do that. you will get less power consumption and stable mining.

[Geraldo]

Are you powering more than two risers from one lane of your PSU?
You should not connect more than two risers to one lane.

SATA: Limited to one per rail.
MOLEX: Limited to two per rail.
6 Pins PCIe: More useful for powering GPUs rather than a Risers.

[somethingiswrong]

Are you powering more than two risers from one lane of your PSU?
You should not connect more than two risers to one lane.

SATA: Limited to one per rail.
MOLEX: Limited to two per rail.
6 Pins PCIe: More useful for powering GPUs rather than a Risers.

Thnanks for the clarification, I was not aware that it's best practice to only use one riser per SATA rail.
The more you know

[psybangas09]

I found my 6 GTX 1070's at 60% power profile (120 core, 720 memory) consumes 760W which on the Silverstone 1000W PSU (1100W peak) will do the same (i.e. restart) occasionally. I only run two riser cards off each SATA power rails and two off molex rail as well. Benefits of gold or platinum rated PSU certainly matters, so when I can afford it I'll like upgrade my PSU to prevent the issue of my mining rig restarting.

[treanski]

Hi guyz!

I have a new mining rig (MSI Z270-A, 8x KFA2 1070Ti, 2x 850W PSU) and have problems running all 8 cards.
When all 8 GPUs are running, rig simply after some 10-20 minutes shuts-off or just stops mining, depending on OS I use.

Settings:
- Above 4G ON, PCIe Gen1 or 2
- No overclocking

Tried:
- With windows 10, SMOS, HiveOS linux...same story
- Exchanged risers...same story
- Changed CPU cooler and thermal paste ....same story
- Changed RAM ....same story
- Underclocking ....same story

Symptoms:
- all 8 GPUs .... stuts-off after 10-20 minutes or simply stops mining
- only 4 GPUs on one PSU    ....works OK
- only 4 GPUs on other PSU .... works OK
- only GPUs 1, 3, 5 and 7 on both PSUs ....works OK
- only GPUs 2, 4, 6 and 8 on both PSUs ....works OK

Any idea?

you are overloading your gpus, they shut down when they overheat. i would guess its the psu which powers mainboard, cpu & risers.
try to undervolt your card or lower clocks or get better psus

[RYXES]

Hi guyz!

I have a new mining rig (MSI Z270-A, 8x KFA2 1070Ti, 2x 850W PSU) and have problems running all 8 cards.
When all 8 GPUs are running, rig simply after some 10-20 minutes shuts-off or just stops mining, depending on OS I use.

Settings:
- Above 4G ON, PCIe Gen1 or 2
- No overclocking

Tried:
- With windows 10, SMOS, HiveOS linux...same story
- Exchanged risers...same story
- Changed CPU cooler and thermal paste ....same story
- Changed RAM ....same story
- Underclocking ....same story

Symptoms:
- all 8 GPUs .... stuts-off after 10-20 minutes or simply stops mining
- only 4 GPUs on one PSU    ....works OK
- only 4 GPUs on other PSU .... works OK
- only GPUs 1, 3, 5 and 7 on both PSUs ....works OK
- only GPUs 2, 4, 6 and 8 on both PSUs ....works OK

Any idea?

you are overloading your gpus, they shut down when they overheat. i would guess its the psu which powers mainboard, cpu & risers.
try to undervolt your card or lower clocks or get better psus

Was just about to say the exact same.

Had the same experience a while back, one of my cables was overloaded and ended up burning the socket.

If it is resetting consider it that your rig is trying to tell you something is wrong with it.

My tip is do not tandem off your cables when possible.

[labsy]

SOLVED!

I tried all your suggestions, but nothing worked.
Then I bought 2nd PCIe-to-4USB adapter, so now ALL risers are connected to those quad-USB adapters (not to single PCIe adapters).
And now it works like a charm!

Seems like my mobo MSI Z270-A Tomahawk is not capable of driving more than 3 GPUS directly via single-PCIe risers.
But Quad-riser did it's job. Weird, but it works.