- Joined
- May 30, 2004
Don't most people use Furmark for OC stability testing?
Isn't the Furmark-throttling like Intel throttling people for using Linpack?
-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
Does Furmark really kill video cards?
- Thread starter SMOKEU
- Start date
Don't most people use Furmark for OC stability testing?
Isn't the Furmark-throttling like Intel throttling people for using Linpack?
Don't most people use Furmark for OC stability testing?
Isn't the Furmark-throttling like Intel throttling people for using Linpack?
Its not the GPU chip its self, its the PCB and the bits that are on it which could get over loaded, PCB's are not designed to run everything screaming in agony, Many AMD and nVidia GPU's have been killed by Furmark.
Its completely unrealistic. A GPU will never do that in the real world.
The same will happen to a Motherboard on P95 if it has weakness.
Its not that GPU PCB's are inherently weak, there not, there just an entirely different thing.
I use benches like 3DMark and Uniqe Heaven to stress test, no game will push a GPU harder. i will never use Furmark.
- Joined
- May 30, 2004
It appeared to have just gotten worse!
It now looks like a power-capping system is in development that will also punish people for using 3D Mark and possibly anything more than granny games!
This looks like something that was possibly made intentionally for the purpose of planned obsolescence..
- Joined
- May 30, 2004
It appeared to have just gotten worse!
It now looks like a power-capping system is in development that will also punish people for using 3D Mark and possibly anything more than granny games!
- Joined
- Jun 22, 2012
I personally still run GPUtool from techpowerup for GPU stability testing, it's pretty much the same thing as furmark but it doesn't give you unreasonable situations to test in.
you could compare GPUtool as the Prime95 of GPUtesting. in the sense that It will give you a worst case scenario situation and will find errors if they are there and you can see what temp your card will top out at, but they will not burn your components un-naturally hard to find the problems.
(unlike furmark or linx)
obviously this is a personal prefence and everyone has their own.
you could compare GPUtool as the Prime95 of GPUtesting. in the sense that It will give you a worst case scenario situation and will find errors if they are there and you can see what temp your card will top out at, but they will not burn your components un-naturally hard to find the problems.
(unlike furmark or linx)
obviously this is a personal prefence and everyone has their own.
Last edited:
- Joined
- May 30, 2004
And IBT is reasonable for CPU cores. Prime95 is inefficient in core testing, sorry.
While Core Damage is not a reasonable test. Core Damage is 32-bit only.
Core Damage don't have error checking.
Prime95 is a waste of time for CPU core testing. I realized that for 10 hours of Prime95, I can find out in 30 minutes!
I don't need to run my PC overnight for Linpack.
While Core Damage is not a reasonable test. Core Damage is 32-bit only.
Core Damage don't have error checking.
Prime95 is a waste of time for CPU core testing. I realized that for 10 hours of Prime95, I can find out in 30 minutes!
I don't need to run my PC overnight for Linpack.
Last edited:
- Joined
- Jun 22, 2012
I will have a look at GPUtool, thanks
Np
it's a really lightweight program as well and requires no installation to the system.don't try to use the overclocking capabilities btw, just use it for stability testing.
And IBT is reasonable for CPU cores. Prime95 is inefficient in core testing, sorry.
Prime95 is a waste of time for CPU core testing. I realized that for 10 hours of Prime95, I can find out in 30 minutes!
I don't need to run my PC overnight for Linpack.
Personally I've never run into any problems if Prime95 couldn't detect them.
10 hours for core testing is not typical. small fft 30mins-1hour is enough to find that out
I typically run it for 6-8hours on blend for a final system stability test. lately even 24. (which imo is totally overkill but i'm waiting for parts to arrive so i might as well kill time making sure.)
Yet on the other hand I've also ran it for 30 minutes on some of my OCs and if they passed that test I never ran into any BSOD or strange behavior.
It's just what works for you mainly I guess.
Last edited:
- Joined
- Dec 7, 2011
What I'm taking from this is that if your 570 survives a few hours of Furmark without releasing the magic smoke, then it can probably survive anything else?
I Furmarked the crap out of my 570 when clocking it -- from the sound of things I got lucky. I've only ever had the one 570 but have heard vanilla 570s like mine have notoriously underbuilt power sections.
I Furmarked the crap out of my 570 when clocking it -- from the sound of things I got lucky. I've only ever had the one 570 but have heard vanilla 570s like mine have notoriously underbuilt power sections.
- Joined
- May 30, 2004
I had flaky OCs with Prime95 failing at 2 hours. So I never gave it the clear until at least a 4 hour run.
- Joined
- Jun 22, 2012
yeah me too apart from that I've never seen a system crash on my final tests beyond the 30 minute marker, by then I typically have the settings dialed down quite tightly
lets try not to hijack the thread too much since it was about gpu stress testing
Last edited:
- Joined
- Jun 10, 2003
- Location
- Plymouth, UK
Using Prime95 & Furmark managed to make my system pull as high as 525 watts at the plug lol
Cutting close to my Seasonic 500w PSU maybe
Running folding@home on GPU and CPU and gaming only pulls 333 watts
-- just felt I should add that I found 10 - 12 hours of Prime 95 (using default Blend selection) on my Q6600 did not find any problems having achieved an apparent stable OC
yet I would crash Fast trying to play FarCry 2 I think it was if I remember right from a few years back. (not folding during this time)
Tweaked my CPU voltage a little more and no more crashing
Happened more than once while I was checking new OC's
Cutting close to my Seasonic 500w PSU maybe
Running folding@home on GPU and CPU and gaming only pulls 333 watts
-- just felt I should add that I found 10 - 12 hours of Prime 95 (using default Blend selection) on my Q6600 did not find any problems having achieved an apparent stable OC
yet I would crash Fast trying to play FarCry 2 I think it was if I remember right from a few years back. (not folding during this time)
Tweaked my CPU voltage a little more and no more crashing
Happened more than once while I was checking new OC's
correct me if im wrong, but if gpu's are being throttled during furmark, doesn't that completely defeat the purpose of using furmark which is to test stability at a given clockspeed?
what am I missing? utter bs imo. If people smoke their cards due to careless ignorance, sucks to be them. All these corporations feel obligated to hold peoples hands for some reason. Blizzard comes to mind.
what am I missing? utter bs imo. If people smoke their cards due to careless ignorance, sucks to be them. All these corporations feel obligated to hold peoples hands for some reason. Blizzard comes to mind.
- Joined
- May 17, 2003
- Location
- South Carolina
Even while throttled it seems to generate more heat (quickly) than Heaven or a typical game. For quickly finding a max overclock and then working out the bugs from there, you can't beat it.
You might have a crash or two in a game and have to fine tune your clocks, but it's great for giving you an idea of where the "sweet spot" is.
I lost TWO gtx 570s to the phenomena described by BobNova earlier in this thread
But while playing a game. Not during furmark.Yes, hand holding ruined MMORPGs. Everquest, too. But we're not all teenagers anymore so some of us need the hand holding these days !
Welcome to the forums, necro klepp !!
- Joined
- May 23, 2008
- Location
- London, England
The only thing Furmark really tests is your cooling system. If your cooling system passes Furmark, it's probably good enough for anything you can throw at it.
magellan
Member
- Joined
- Jul 20, 2002
Yep and falling doesn't kill you, it's the abrupt stopping at the end that does.
- Joined
- Jun 9, 2013
Deceleration trauma.
I got the Gigabyte Radeon 280X rev. 2. When I run Furmark with +20% powertarget, my system just shuts down and boots again after 3 sec. I know this might be due to a insufficient power supply, but mine should handle it well. It is a "be quiet! Straight Power E9-CM 580W". Is it possible, that the card draws so much power? Below are my furmark settings:
GPU temperature never was above 75° C.
I then tested the card in heaven benchmark and everything was fine. Could it be damaged?? How would I notice it??
GPU temperature never was above 75° C.
I then tested the card in heaven benchmark and everything was fine. Could it be damaged?? How would I notice it??
Similar threads
