GTX 970 causing restarts during benchmarking

[Chumpster]

Current specs:
AMD Phenom II BE 965 OC to 3.8 GHZ
Asus m5A97 ATX
Corsair xms3 ddr3 (2x2)
FD Newton R2 650W 80+ (It's 6 years old)
EVGA GeForce GTX 970 SSC ACX 2.0

I had an old AMD HD6870 graphics card running in the system and it never caused restart during benchmarking (And rechecked today and it worked fine).

So I bought this card second hand yesterday. I installed it without any problems and ran some benchmarks. At first, I tried unigine Superposition benchmark multiple times, it rebooted during the same scene every time. Then I tried unigine Valley and it rebooted during the same scene. All 3dmark benchmarks rebooted on me. I messed with it in afterburner and Valley and Superposition then passed all scenes, but the 3dmark benchmarks still crash.

I ran AIDA64, no problems. 
12V voltages show 12.107 - 12.158 on HWMonitor. 
I ran Prime95 and memtest as well, and everything was stable. 
I've also tried different BIOS versions.

Unfortunately all my friends are out of town so I can't test the GPU with another PSU yet. What do you guys think? Is it the PSU or GPU?  

[RadiatingLight]

Probably the GPU.

try bumping up the voltage a bit, see if that's the issue.

[Chumpster]

Just now, RadiatingLight said:

Probably the GPU.

try bumping up the voltage a bit, see if that's the issue.

Core Voltage on afterburner? 

[m0k]

4 minutes ago, Chumpster said:

So I bought this card second hand yesterday. I installed it without any problems and ran some benchmarks

just wanted to be 100% clear
did you use DDU to remove the old AMD 6870 drivers before installing the nvidia drivers?

[Chumpster]

Just now, mok said:

just wanted to be 100% clear
did you use DDU to remove the old AMD 6870 drivers before installing the nvidia drivers?

No, I used AMDs official uninstaller.

[Ryujin2003]

What critical error does Event Viewer tell you? If it's NVIDEAKRNL or something like that, it's GPU related. Some random powerkrnl unknown shutdown, I'd look at PSU.

[RadiatingLight]

7 minutes ago, Chumpster said:

No, I used AMDs official uninstaller.

use DDU and try again. 

[Chumpster]

Just now, Ryujin2003 said:

What critical error does Event Viewer tell you? If it's NVIDEAKRNL or something like that, it's GPU related. Some random powerkrnl unknown shutdown, I'd look at PSU.

I've checked and all critical error comes from KERNEL-POWER

 

Just now, RadiatingLight said:

use DDU and try again. 

Done, Now I've gotten a bit further into the Firestrike benchmark, but it still rebooted.

[RadiatingLight]

Just now, Chumpster said:

I've checked and all critical error comes from KERNEL-POWER

 

Done, Now I've gotten a bit further into the Firestrike benchmark, but it still rebooted.

have you tried raising the power limit and giving it a bit more voltage in afterburner.

[Chumpster]

Just now, RadiatingLight said:

have you tried raising the power limit and giving it a bit more voltage in afterburner.

Never OC'd it before, how much would you recommend?

[RadiatingLight]

6 minutes ago, Chumpster said:

Never OC'd it before, how much would you recommend?

nothing in there is dangerous, and I personally just crank everything up to maximum.

if you want to be a bit more cautious, you can go full power limit, and not go all the way with voltage

[Chumpster]

Just now, RadiatingLight said:

nothing in there is dangerous, and I personally just crank everything up to maximum.

if you want to be a bit more cautious, you can go full power limit, and not go all the way with voltage

Cranked voltage up halfway (20mV), power limit all the way. Frequency up a bit etc etc. Now it reboots during an earlier frame. 

[RadiatingLight]

1 hour ago, Chumpster said:

Cranked voltage up halfway (20mV), power limit all the way. Frequency up a bit etc etc. Now it reboots during an earlier frame. 

might just be broken.

sorry.

could be psu though, so test in a friend's sytem if possible.