[Archer S]

First and foremost, full system specs as its for a friend, not my computer.

i7 950 [email protected] 3.8GHz
12GB (3x4GB) DDR3 @1600MHz
Gigabyte X58A-UD3R
1kW PSU Cougar CMX
3 x EVGA GTX 570 HD 1.25Gb GDDR5
2x 128GB SSD (RAID0)
3x 1TB HDD (no raid) (WD green, WD Black, Seagate barracuda)

Now for the problem

Upon power, the machine posts and gets as far as the windows logo after which the monitors go black saying no signal and the system hangs until a forced reset.

The troubleshooting that has been done so far
Tried every card individually with one monitor, same issue no changes
Tried every stick of ram individually in all slots, same issue
Cleared CMOS returned all settings to stock, same issue
Underclocked the cpu, no changes, same issue
Removed all hard drives except primary raid setup, no change
Fresh install of windows (SSDs formatted):

-computer worked and ran off the vga default drivers. after installing the nvidia drivers the computer would boot and work normal for about 15 minutes after which the system would hang in the same manner and never boot again with the same issue.
-works fine in safe mode though

Computer works fine with a different graphics card (some workstation ATI card), however these 570s have yet to be tested in a different rig.

Suspect is the data path from video cards to cpu. perhaps chipset issue? Anyone have any ideas what could be happening here or any other troubleshooting tips that could shed some light on it? Motherboard or vga problem?

 

[cgipson1]

Sounds like either a driver, or a power issue (more driver).

Did you try an earlier version of the drivers? (assuming you installed the latest)?

Did you try a single card (non-SLI) on the fresh install, or was it all three cards again?

Is this a new build, or have the components been proven elsewhere?

 

[Archer S]

The components have been running for years, and the cards have been mining doge 24/7 for about a month and a half?

The fresh install was with the 3 cards together. Multiple driver versions have been tried not just the latest

 

[cgipson1]

Quote:

Originally Posted by Archer S

The components have been running for years, and the cards have been mining doge 24/7 for about a month and a half?

The fresh install was with the 3 cards together. Multiple driver versions have been tried not just the latest

Hmmm... hard to make a call on that.

Test the GPU's in a different system obviously... maybe double check the PSU... see if it is no longer able to provide adequate juice for those three cards? Might be chipset or something else on the motherboard... that chokes thermally once it hits a certain stress point...

but it is all conjecture... only more testing will help isolate the issue. That is about all I have....

 

[Archer S]

Quote:

Originally Posted by cgipson1

Hmmm... hard to make a call on that.

Test the GPU's in a different system obviously... maybe double check the PSU... see if it is no longer able to provide adequate juice for those three cards? Might be chipset or something else on the motherboard... that chokes thermally once it hits a certain stress point...

but it is all conjecture... only more testing will help isolate the issue. That is about I all I have....

Well the PSU would not power even 1 card individually. So i doubt its the issue but, acknowledged, will check. The freezing happens at pretty much idle.

Also i was mistaken, it only mined 12/7. but i doubt thats relevant.

Either way thanks for your advice

 

[cgipson1]

Quote:

Originally Posted by Archer S

Well the PSU would not power even 1 card individually. So i doubt its the issue but, acknowledged, will check. The freezing happens at pretty much idle.

Also i was mistaken, it only mined 12/7. but i doubt thats relevant.

Either way thanks for your advice

anytime... good luck with it. Let us know what you find...

 

[ar3f]

Did you check "Event Manager" (in Admin Tools) for errors?

 

[cgipson1]

Quote:

Originally Posted by ar3f

Did you check "Event Manager" (in Admin Tools) for errors?

Good idea.... might even try to read a dump, if there is one being generated...