New PC Freezing During Anything GPU Intensive

Citizen Burns

Distinguished
Mar 20, 2011
7
0
18,510
So I purchased a custom PC from magicmicro.com and they did a pretty standup job. The inside is nice and tidy and all in all I'm pleased. On the other hand... the PC does not play well on any game. I have my specs from CPU-Z in my sig but I'll pop it right here as well. Let it be known that I do not OC and everything listed below is at stock settings.

Mobo: ASRock 870 Extreme3
CPU: 3.5Ghz Phenom X4 970 (V6 Coolermaster dual fan with several heatsinks keeping my CPU frosty)
GPU: HD Radeon 6970
RAM: 8 Gigs DDR3 1333 (2x4)
PSU: 750w Thermaltake
Cooling: 5 case fans (120mm front, 200mm side, 80mm bottom intake) (120mm rear, 200mm top exhaust)
OS: Win 7 64 bit
HDD: 1 TB Western Digital Black

So to start the first system I received I had to end up returning for an exact copy. I was having the problem of course on the original system but I could at least play for 4-6 hours sometimes before I froze up. I tried a few things before I sent it back, I replaced both sticks of RAM in the original system and the problem stuck around. I tried different RAM ports (All of them in fact) and the problem would not dissappear. At that point I didn't want to experiment anymore and sent it back for a replacement.

So the new one gets here yesterday and to my surprise the problem is not only present in the new PC but I can't even get a game to run for 5 minutes. Of course I'm rather upset at the circumstances but I'm holding my head high and turning to the community to maybe offer suggestions of where to turn next. Everything else on my system runs fluid, I can surf, download, stream music and video with no problems. My PC will withstand Furmark for roughly 3-4 minutes before it decides to freeze up as well so it isn't the games I am playing. Memtest86 runs for 24+ hours with no issues as well.

Lastly, I have checked several times on the heat in my PC because it scared me at first (Idling around 40-45c and max load around 75-85c.) AMD shed some light on the situation by letting me know that they didn't consider the card hot until around 125c so learning that was both relieving and troubling (125c?!?!)

Anywho, here are my hunches so far:

1: Mobo compatability (Though I haven't found any conflicts based on ASRock's specs on their site for my mobo.)
2: GPU needs RMA'd (Also skeptical of this one because this is the second GPU I have had with the issue.)
3: Not enough power? (I was assured 750w was more than enough for a single 6970 and all the components I have.)
4: Ninja RAM (Like I said Memtest86 says my RAM rocks but I have heard stories of bad RAM that ninjas testing.)
5: RAM timings? (I have no clue how to do this myself but maybe the 1333 RAM isn't meshing?)

Anyone have any thoughts on the above? Or possibly something I haven't thought of because I am nowhere near an expert. I would appreciate any thing you guys and gals have to offer.
 

pacioli

Distinguished
Nov 22, 2010
1,040
0
19,360
125 C is not acceptable... Water boils at 100 C
85C for a CPU is very hot. It is CPU killing hot!
Do you have temperatures for your GPU?
It sounds like you may have a power supply issue. 750 watts should be more than plenty and Thermaltake is considered to be a reputable brand but sometimes you just get a bad one. Sometimes you get two bad ones...
 

thenewnumber2

Distinguished
Jan 14, 2009
275
2
18,815
Your CPU is far from "frosty": 75-85C under load is way too hot. No surprise you're crashing... I think you'll find that your CPU heatsink isn't properly attached, or they didn't bother with the thermal compound. That's the first thing to sort out.

The New Number Two
 

Citizen Burns

Distinguished
Mar 20, 2011
7
0
18,510


Its not my CPU running at that temp thats my GPU. MY CPU is running at 40-45c under load. The highest I have ever gotten my CPU is 50c after an hour long Prime95 run. I dont have any heat issues in this rig, I know it is usually one of the most popular reasons for someone to crash but my system is not overheating. Thanks for the input though ;)
 

Citizen Burns

Distinguished
Mar 20, 2011
7
0
18,510


I must not type very clearly lol :D

My GPU isn't running at that temp, it is nowhere close to that in fact. The highest my system has ever hit is around 85c under load and the 6970 is definately built to handle that. Newer cards are built to handle temps that would have made us cringe 2-4 years ago with older models. If I recall correctly it has something to do with a "vapor chamber" but I have no clue what this is or even means. All I know is that the problems I am having are not heat related. Thanks for the input though :sol:
 

pacioli

Distinguished
Nov 22, 2010
1,040
0
19,360
The AMD Radeon 6970 Does not have a vapor chamber. Nvidia GTX 570 and 580 are the only two cards with a vapor chamber. They basically have a liquid that boils off and re-condenses in the heatsink.

I still think you are running on the hot side for a GPU... I don't care what AMD says... 125C is too hot. A hot GPU will give you artifacts and freezes in games.

Your problem seems to be possibly related to a bad power supply or insufficient power. 750 watts should be plenty of juice but if the PSU is not putting that much wattage out you wil have problems.