xfire 7950s running too hot?

N

n370zed

Limp Gawd
Joined
May 21, 2012
Messages
450
I recently got these http://www.newegg.com/Product/Product.aspx?Item=N82E16814127737 and I can't game for more than 10-15 before the PC shuts down and starts back up again on its own. I cracked open my case which is a HAF 922 with a 200mm side fan and the 2nd card is very hot. Ive noticed the 2nd card is always hotter than the top card.I have a Kingwin 750w PSU and the 7950s are stock. Cards were bought last Friday.
 
Are you running a program to monitor the temps on your cards? With the HAF 922, I would be surprised if it is purely a heat issue.

Also, 10-15 what? Seconds, minutes?
 
MrC4 said:
Are you running a program to monitor the temps on your cards? With the HAF 922, I would be surprised if it is purely a heat issue.

Also, 10-15 what? Seconds, minutes?
Click to expand...

Sorry 10 to 15 minutes. I'm using MSI afterburner to monitor temps. I can do a full run of 3dmark11 and temps were 72c. But when I tried BF4 and Sleeping Dog PC shuts down and starts up again. My PSU has only a 6 and 6+2 pin. I had to use 2 molex to 6 pin adapters for the 2nd card.
 
Last edited:
I have those cards in Crossfire, I just bought them last week.

They overheat, my top card easily reaches 100+ Celcius. I have no space between them because my mobo is mATX.

I want to return them and get something with a reference cooler but I have already used the AMD game code things and newegg says they are $99 each. :rolleyes:
 
Single fan card in a smaller 922 could be an issue.

Disable ULPS and go into the settings and you can make it display the second GPU's temp.

Also, how are the cards stacked? Any space between them at all?
 
n370zed said:
Sorry 10 to 15 minutes. I'm using MSI afterburner to monitor temps. I can do a full run of 3dmark11 and temps were 72c. But when I tried BF4 and Sleeping Dog PC shuts down and starts up again. My PSU has only a 6 and 6+2 pin. I had to use 2 miles to 6 pin adapters for the 2nd card.
Click to expand...

3D Mark does not stress the cards because each test is only like 30 seconds long, then the card gets to cool back down while its loading the next test.

My cards dont overheat on 3DMark either, even when overclocked to 1150mhz core.
 
If you run the cards individually, what are the temps? Single fan + tight crossfire sounds like they are air starved. I'd say the top fan is picking up the hot air right behind the gpu of the bottom card.
 
Priller said:
Single fan card in a smaller 922 could be an issue.

Disable ULPS and go into the settings and you can make it display the second GPU's temp.

Also, how are the cards stacked? Any space between them at all?
Click to expand...

I can easily slide my hand between the cards. I did the ULPS last night but it didn't help.
 
leeleatherwood said:
I have those cards in Crossfire, I just bought them last week.

They overheat, my top card easily reaches 100+ Celcius. I have no space between them because my mobo is mATX.

I want to return them and get something with a reference cooler but I have already used the AMD game code things and newegg says they are $99 each. :rolleyes:
Click to expand...

Here's the thing tho my bottom card is hot as hell because I can feel the back end of where the core and metal backing are. The top card is warm but not unstrikingly hot as the bottom card.
 
n370zed said:
Here's the thing tho my bottom card is hot as hell because I can feel the back end of where the core and metal backing are. The top card is warm but not unstrikingly hot as the bottom card.
Click to expand...

Check the voltage.

For some reason my cards came with default voltage of 1250 (very high). I can set them at 1050 and they run fine, with much lower temps.

My top one still overheats though, lack of airflow :(
 
Try uninstalling msi afterburner. If one of the cards has a funky BIOS on it AB may be ramping the voltage up.

Test the hotter card by itself. Normally the top card would be hotter, no?

That card maybe has a VRM temperature problem. Use HWinFO64 to check out the temp sensors on your VRMS. If they get towards 100C there's a big problem considering you're at stock speeds. Shouldn't go above 80 at stock, I should think.

I like to monitor render temps with Unigene Valley.
Where did you get the cards? Shrinkwrapped?

Those cards should be set at 1.15v iirc, if one of the BIOS got tinkered with you can edit it. They may be voltage locked, if so only editing the BIOS will change the voltage as far as I know so far. Even if MSI AB is telling you you changed the voltage, it may not at all. Again you can check this with HWMonitor or HWinFO64. Been working with Powercolor 7950 BIOS options lately myself. That was my issue. 1.25v vs. 1.15v after a BIOS edit to allow power range 50% instead of 20%.
 
Bluesun311 said:
Try uninstalling msi afterburner. If one of the cards has a funky BIOS on it AB may be ramping the voltage up.

Test the hotter card by itself. Normally the top card would be hotter, no?

That card maybe has a VRM temperature problem. Use HWinFO64 to check out the temp sensors on your VRMS. If they get towards 100C there's a big problem considering you're at stock speeds. Shouldn't go above 80 at stock, I should think.

I like to monitor render temps with Unigene Valley.
Where did you get the cards? Shrinkwrapped?
Click to expand...

Newegg and no the box the cards came in were new but no shrink wrap. And the card is at 1.25v at stock which is 840 MHz boost. BTW I'm at work and won't be able to do anything until tonight. But thank you for the help so far. Please keep them coming. I was playing BF at 1920x1200 ultra with 2xmsaa and I was seeing 80-100 fps and it was so smooth.
 
n370zed said:
Newegg and no the box the cards came in were new but no shrink wrap. And the card is at 1.25v at stock which is 840 MHz boost.
Click to expand...

Yeah that voltage is insane for a stock card.

Google the stock voltage of other 7950's and you will see almost all of them are below 1100.
 
leeleatherwood said:
Yeah that voltage is insane for a stock card.

Google the stock voltage of other 7950's and you will see almost all of them are below 1100.
Click to expand...

Thanks I will definitely undervolt it to 1.10 and hope it will help. I don't plan on overclocking because its pointless with the performance from 2 GPUs.
 
Bluesun311 said:
Try uninstalling msi afterburner. If one of the cards has a funky BIOS on it AB may be ramping the voltage up.

Test the hotter card by itself. Normally the top card would be hotter, no?

That card maybe has a VRM temperature problem. Use HWinFO64 to check out the temp sensors on your VRMS. If they get towards 100C there's a big problem considering you're at stock speeds. Shouldn't go above 80 at stock, I should think.

I like to monitor render temps with Unigene Valley.
Where did you get the cards? Shrinkwrapped?

Those cards should be set at 1.15v iirc, if one of the BIOS got tinkered with you can edit it. They may be voltage locked, if so only editing the BIOS will change the voltage as far as I know so far. Even if MSI AB is telling you you changed the voltage, it may not at all. Again you can check this with HWMonitor or HWinFO64. Been working with Powercolor 7950 BIOS options lately myself. That was my issue. 1.25v vs. 1.15v after a BIOS edit to allow power range 50% instead of 20%.
Click to expand...

Thank you for this. I'm feeling much better now. There's some light ahead.
 
I removed the shroud on both of my cards and my temps dropped 25c+

You can remove the shroud without voiding the warranty, once removed you can tell from the design of the shroud that all it does is trap in heat, its a horrible design, the worst I have ever seen.

Anyways, I am now running 1000mhz/1500, 1100vcore and temps do not exceed 80c, even when directly next to to each other. (no gap in between)

UPDATE: 1200/1600mhz, 1250vcore, max temp 95c in crossfire with the shrouds removed. No crashing!
 
Last edited:
leeleatherwood said:
I removed the shroud on both of my cards and my temps dropped 25c+

You can remove the shroud without voiding the warranty, once removed you can tell from the design of the shroud that all it does is trap in heat, its a horrible design, the worst I have ever seen.

Anyways, I am now running 1000mhz/1500, 1100vcore and temps do not exceed 80c, even when directly next to to each other. (no gap in between)

UPDATE: 1200/1600mhz, 1250vcore, max temp 95c in crossfire with the shrouds removed. No crashing!
Click to expand...

I did just that. I removed the POS plastic shroud. I was gonna apply some AS5 but then I saw the littel sticker on the back about voiding the warranty. Heres the thing. I tried stock at 1.15v. I tried sleeping dogs and pc crashed when one of the GPU hit 85c.Took abot 5 minutes of gameplay. I did a fresh install of Windows. I installed Trixx instead of AFterburner and still my system crashes. Im at a loss :( im gonna try playing with a single card is see what happens.
 
I run my 7950's at 1.05v. My 3.3v on my power supply is sagging to 3.2v and drops to 3.15v when I start playing games. All other voltages are perfect. If I try to max out Sleeping Dogs or Tomb Raider it will eventually restart the PC. Everything else runs perfectly. I have a 850w CoolerMaster power supply.

Now to put this into perspective, if I just run my original Sapphire 7950, I can feed it 1.2v without problems. I need a new power supply. I would surmise that you do also.

My Sapphire 7950 will restart my PC if it hits 95c.

:)
 
Oh... Yah maybe I should have looked at that b4 I said anything. I just joined the xfire club today myself. After I decided I was going to go xfire the first thing I did was save up for a Seasonic platinum1000. I decided to be weird and get a 280x and see if it would work with my Powercolor 7950 V5E that I put a Gelid icy vision A on.

I dunno why but I had to put the 280x on the bottom, but it works! After an hour of unigine valley it's hot but I don't smell anything burning. 78 on the top 71 on the bottom with the case closed but we've got it at like 74F so pretty low ambient. I think the Gelid would do better but the fan blades are literally almost touching the back of the 280x. Kinda bizarre - had to plug the monitor into the 280x or no xfire and it had to be the bottom card. Took a while to troubleshoot...

Anyway 600W is cutting it close but u may also have 1 card with a bad component.
 
I think I figured out the problem. Sleeping Dogs was the only game so far that crashes my PC. I tried Batman AC, BF4, Skyrim, and The Witcher 2 and they all ran just fine. I'm at 950/1250 1.15v. The temps on the vrm prior to removing the plastic shrouds were hitting 98c now its in the 60s. I'm loving the super smooth gameplay of having a dual GPU setup. Now does anyone know where I can get some VRM sinks because they are still running pretty warm.
 
Sleeping Dogs is one of the premier games I use to test the stability of my system. Try monitoring your voltage to see if you're having problems with it. If it's a heat issue try lowering your voltage just a tad more. You should be able to OC your cards much higher so 950 is easy to do at 1.05v.
 
Well, the ones I used came with the Gelid Icy Vision A package. They would probably work really well on yours because it is a reference pcb. The vrms get pretty hot and resist cheap thermal tape so gelid has you screw in one long heatsink with thermal paste on to cover the vrms near the power inputs. My pcb did not fit this solution. So I used regular heatsinks and hq thermal pad. AL you really need to is get some hq thermal pad to place between the vrms and the stock heatsink I imagine.
 
It was indeed the power supply. The old 750 psu was giving my cards1.25v. The new Rosewill 80+ Platinum was showing 1.15v. I'm running 1000mhz/1500 stock and am loving the performance. I was fooled into buying a PSU based on advertised wattage. Lesson learned.
 
You must log in or register to reply here.
Back
Top