Getting BSOD, every second day, BSOD REASON:kmode exception not handled, Testing MEMTEST86. have some Q's.

[bug]

Guys, i just disabled my 3 gpus out of 5, what i found is theres no more lag, what can be the issue? Performance and stability wise i tested all one by one each, all are working perfect and stable..
this is new issue ive seen

Monitor the cards and figure it out for yourself? Check voltages, running frequencies, temps... the works.

 

[Honey]

Possibly your 5 x GTX1070 are not being handled by your PSU? Maybe just a case that the CPU can't push out all of that grunt on 4 cores alone, it could be a number of things.

Really? But i saw people using celeron most of the time with multiple gpu Actually running 4 gpus on one psu that is thermaltake toughpower 1000w gold, and one card by corsair 500w psu

Monitor the cards and figure it out for yourself? Check voltages, running frequencies, temps... the works.

j never seen this thing before, im running these settings from year, no stability issue, since im getting bsod and stutter, ive restored default the settings.

 

[cadaveca]

Possibly your 5 x GTX1070 are not being handled by your PSU? Maybe just a case that the CPU can't push out all of that grunt on 4 cores alone, it could be a number of things.

Yes, my first thought and approach with issues like this is the PSU. VGAs do weird things at idle, and the smallest thing can cause a jump to higher clocks from idle and push a psu that isn’t quite capable

Really? But i saw people using celeron most of the time with multiple gpu Actually running 4 gpus on one psu that is thermaltake toughpower 1000w gold, and one card by corsair 500w psu

Yeah those four gpus are likely a bit too much for the psu. It could be something else for sure, but crash on idle is usually memory controller issues or inadequate power supply. Since memory seems to test fine, that does point at the psu, but of course there is a chance it could be something else.

 

[bug]

j never seen this thing before, im running these settings from year, no stability issue, since im getting bsod and stutter, ive restored default the settings.

Something's giving up the ghost in that box. You need to inspect one component at a time and figure out what. We can't do that over here.
So far, you've got suggestions for checking memory, video cards and PSU. Start going through that list.

 

[Caring1]

Disconnect one GPU from the PSU that has four connected, then see if the fault continues.

 

[Honey]

4 gtx 1070 cant run by 1000w PSU?

 

[bug]

4 gtx 1070 cant run by 1000w PSU?

Do you remember the PSU is one of the components that could potentially be faulty?

 

[cadaveca]

4 gtx 1070 cant run by 1000w PSU?

For me personally, no, that is not enough. I run my 2x 1080 on a 1200W PSU with X299 and X399 CPUs, and will go no lower than 1000W for Z370/X370. Part of this is because I came from 1070's, where each GPU can pull around 200W, or a bit more. I am likely only pulling <750W from that wall with my entire system, but once you get close to the limit of the PSU, the voltage it supplies may droop below acceptable levels. With 4 VGAs on 1000W PSU, that's 800 W for the VGAs, plus the rest of the system. Could be pushing that PSU, yep.

 

[erpguy53]

I saw the specs of the Thermaltake Toughpower 1000w PSU
maybe OP needs to get a Toughpower 1200w though that one may be a little bigger (in size) than the 1000w PSU.
and since OP also uses an HD & SSD (which also uses power on the 12v rails along with the GTX 1070 cards), consider replacing the old HD with an external USB hard drive that can be connected to a USB 2.0/3.0 port.

 

[hat]

Now that I think of it, this issue seems oddly familiar to one I was experiencing not too long ago. Are you using powered risers for those cards? If not, those cards are still drawing a lot of power through the slots, and the 24-pin isn't keeping up. If that's the case, both the connector on the board and the 24-pin power supply connector could be burned up where the two +12v wires are, and the wires themselves. This is the reason why all of those crazy mining motherboards with huge amounts of PCI-E slots have a bunch of extra PCI-E, molex, and even 24-pin power connectors. Of course, that can be avoided by using risers that take their own power source, such as these.