Safe to run furmark overnight?

[straxusii]

I wouldnt, I think it's fair to say furmark can stress the parts other apps dont reach!

 

[Raves]

no

 

[unknownsock]

A bit of research would have told you that the hot V-REG with furmark is associated with the 4850/4870 cards not the 5000 cards, so the problem is not associated with all ATI cards.

I thought this was also brand specific, like Powercolor too?

 

[buttkinz]

I would do as the chap above, loop a benchmark such as heaven for a couple of hours, don't forget motherboards can regularly go wrong as well lol

 

[wolvers]

IT does still put an unnatural load on the v-reg's on a 5series

I found that it only roasts the VDDCs if you don't apply high levels of AA. Have a try yourself and see.

 

[Phil2008]

Normal gaming my 3870 get to about the 75c mark, but running furmark it gets to 92c. So furmark puts a hell of a lot more strain on your gpu, then normal gaming does

 

[Admiral Huddy]

furmark puts a hell of a lot more strain on your gpu, then normal gaming does

Source?



So are you all saying these cards have never been lab tested for overheating on stress before launch? Interesting thoughts.

To be honest, I'd just run a dozen or so loops of Crysis and STALKER CoP Benchmarks rather than futuremark.

 

[mmj_uk]

In a properly designed GPU there is no harm in running Furmark for prolonged periods other than a little extra wear and tear, the 4x00 series debacle was simply a result of AMD cutting corners in the power circuitry and being caught out, they have since rectified the situation.

Motherboards and practically every other GPU ever made were designed with "unnatural loads" (ie. very efficient use) in mind.

 

[el Kuarlos]

Its the equivalent of telling a guy who had a pain in his chest to run three marathons back to back.

Yea if he gets through he probably doesn't have a dodgy ticker...
but if he drops down dead or it gets worse, it could just as likley be you told him to do something that he would never ordinarily do way above and beyond real world stress levels.

There are better ways of addressing instability issues like the one you have...furmark isnt like a cpu stress test, it creates a truly abnormal load and I wouldn't want to use it for anything other than discovering worst worst case temperatures or benchmarking over short periods.

Quote:
Originally Posted by Phil2008 View Post
furmark puts a hell of a lot more strain on your gpu, then normal gaming does

Source?

Simple power consumption and temperature alongside the fact overclocks that fail under furmark often either dont fail in some games or will last a lot longer- why would you need a source to deduce this?

 

[mithrandir]

I personally wouldn't run furmark overnight but I have generally found once it has passed 'x' minutes it tails off and temps will remain constant.

If you're wanting to test gpu stability, I would instead loop a demo bench such as DM4 or as mentioned above an unigine benchy.

 

[SirCanealot]

Sounds like a complete waste of power to me...

I run ATITools' artifact scanner most of the time, but I feel even that is a bit too harsh, and that runs about 10c cooler than furmark!

The most I ever stress tested was a 9 hour Prime 95 session. And that was only because I was about to start running WCG and I wanted to make sure I wouldn't be sending back bad data...

 

[Phil2008]

Source?



So are you all saying these cards have never been lab tested for overheating on stress before launch? Interesting thoughts.

To be honest, I'd just run a dozen or so loops of Crysis and STALKER CoP Benchmarks rather than futuremark.

Well when I just quit crysis the gpu temp is 65 - 70c

 

[JTrickle]

Source?

He doesnt need one for that statement. It stands to reason/common knowledge.

I want to test my gpu and wondered if its safe to run furmark overnight or while I'm at work?

If you suspect a problem with your gpu, just try it for 10-15 mins. Ive seen gfx issues that furmark didnt reveal, but it doesnt often fail to spot them.

 

[Ordokai]

Simply no. There's a very specific point at which the temps and fan speeds will just stay constant, I bet this point can be achieved in like an hour. There is no need to test it overnight, also if you're looking for stability/artefacts furmark is not the best software to use. Basically you will come back in the morning and learn nothing.



- Ordokai

 

[ferret face]

but surely if it crashes then the card is faulty?

 

[Raves]

as mentioned previously : no

 

[el Kuarlos]

Actually having had a closer look, running furmark overnight has changed from a pointless idea to a bad one in my eyes...

http://www.geeks3d.com/20090925/ati...on-against-power-virus-like-furmark-and-occt/

http://www.geeks3d.com/20090916/furmark-slowdown-by-catalyst-graphics-drivers-is-intentional/

 

[Admiral Huddy]

Its the equivalent of telling a guy who had a pain in his chest to run three marathons back to back.

haha made my day this comment!

 

[infernox]

your card will melt if you do it overnight.. so i wouldnt