Jump to content

GPU-fans stop working


Recommended Posts

Hi,

I'm looking for a tool to test my GPU.

Lately I've had my fans stop halfway through a mission. In my temperature monitoring is see the fan-speed-% indicactor stay at 80% or higher and the fan-speed-rpm indicators go to zero (fans effectively stop turning), while the temperature keeps climbing, the fan-speed-% indicator climbs towards 100% (rpm stays zero) until the GPU starts throttling back and the game goes from +100 fps down to +-10fps. 

The real problem is, is that it doesn't happen every time. I cannot reproduce it reliably. I think it's a hardware problem, because if it was a software or a settings-problem it would happen each time, which it doesn't.

That's why I'd like to use some kind of stress tool to put the gpu under stress and perhaps see what's causing this problem. 

I've updated to the latest driver, but that didn't solve the problem.

Regards

 

i7 6700k/GTX1070-8G/MSI-Z170A Gaming Pro Carbon/32GB DDR4 Kingston HyperX PREDATOR DDR4 3000MHZ Vengeance 1600/TM Warthog #6106/Samsung SB350_S27B350H/OCZ Agility3 SSD 128GB / Win10-64/TIR5



 

 

Link to comment
Share on other sites

1 hour ago, bengo said:

I think it's a hardware problem, because if it was a software or a settings-problem it would happen each time, which it doesn't.

No disrespect intended, but that's a largely misguided line of reasoning.  Software people (game developers and programmers) tend to say that, but it's absolutely not reality.

Hardware people say the same thing sometimes, and they're no more correct than the software people.  It's just not a concrete fact, either way.

For one thing, you can't say with any real certainty that the software is seeing the exact same set of conditions every time and thus always responds the same.  It could be seeing factor(s) you're unable to monitor and therefore it will behave differently without your seeing an apparent reason.

Another thing is that software can absolutely have 'bugs' that do not occur predictably and reliably with the exact same apparent conditions.

I've worked in computerized systems maintenance for over 40 years, and I can promise you that neither software nor hardware is bound to the 'rule' you're trying to apply.

That said, I might be inclined to agree with your conclusion, but not for the reason you've cited, and not without further details 🙂  (among them, where your measurement data is coming from).

Something else worth considering: Are you including the GPU BIOS among 'software'?  That's what it is, and there have been instances where it caused fans to misbehave.  So if you haven't thought about that, it might not be a bad idea.  You don't mention your GPU manufacturer, but some use Windows utilities that can update the GPU BIOS, and it might've been changed without your necessarily realizing what happened.

As far as utilities go, I use 3DMark FireStrike; it's a reasonable stress test - there are others, of course.  But 3DMark in it's basic form is free, is well-known and highly regarded, and it's safe for testing - arguably, some other tools are not as safe (FurMark comes to mind here).


Edited by kksnowbear
  • Like 2

Free professional advice: Do not rely upon any advice concerning computers from anyone who uses the terms "beast" or "rocking" to refer to computer hardware.  Just...don't.  You've been warned.

While we're at it, people should stop using the term "uplift" to convey "increase".  This is a technical endeavor, we're not in church or at the movies - and it's science, not drama.

Link to comment
Share on other sites

1 hour ago, bengo said:

I'm looking for a tool to test my GPU.

 

48 minutes ago, kksnowbear said:

3DMark FireStrike

this will definitely make your fans spin. if they don't then you have a hardware issue. the utility is supposed to stress test all hardware.

  • Like 2

AKA_SilverDevil AKA Forums My YouTube

“It is better to keep your mouth closed and let people think you are a fool than to open it and remove all doubt.” — Mark Twain

Link to comment
Share on other sites

Okay, seems this 3DMark only is available through Steam, which I don't have installed. Any other tools recommended ?

 

i7 6700k/GTX1070-8G/MSI-Z170A Gaming Pro Carbon/32GB DDR4 Kingston HyperX PREDATOR DDR4 3000MHZ Vengeance 1600/TM Warthog #6106/Samsung SB350_S27B350H/OCZ Agility3 SSD 128GB / Win10-64/TIR5



 

 

Link to comment
Share on other sites

24 minutes ago, bengo said:

Okay, seems this 3DMark only is available through Steam, which I don't have installed. Any other tools recommended ?

 

i got it from there too. it used to have a free trial version. honestly i do not know any others. this is the de-facto ruler of the testing utils.

AKA_SilverDevil AKA Forums My YouTube

“It is better to keep your mouth closed and let people think you are a fool than to open it and remove all doubt.” — Mark Twain

Link to comment
Share on other sites

17 hours ago, bengo said:

Any other tools recommended ?

Cinebench perhaps?
https://www.maxon.net/en/cinebench

  • Like 1

Win11 Pro 64-bit, Ryzen 5800X3D, Corsair H115i, Gigabyte X570S UD, EVGA 3080Ti XC3 Ultra 12GB, 64 GB DDR4 G.Skill 3600. Monitors: LG 27GL850-B27 2560x1440 + Samsung SyncMaster 2443 1920x1200, HOTAS: Warthog with Virpil WarBRD base, MFG Crosswind combat pedals, TrackIR4, Rift-S.

Personal Wish List: A6 Intruder, Vietnam theater, decent ATC module, better VR performance!

Link to comment
Share on other sites

5 minutes ago, Lange_666 said:

Cinebench stresses only the CPU.

@bengo

But Heaven Benchmark is free afaik. It's a good approximation for DCS, since it is also DX11. I use it all the time.

 

https://www.guru3d.com/download/download-unigine-heaven-benchmark/


Edited by Hiob
  • Like 1

"Muß ich denn jedes Mal, wenn ich sauge oder saugblase den Schlauchstecker in die Schlauchnut schieben?"

Link to comment
Share on other sites

2 hours ago, Hiob said:

But Heaven Benchmark is free afaik. It's a good approximation for DCS, since it is also DX11. I use it all the time.

Of course; there are several free alternatives I'm aware of.

Problem with the Unigine benchmarks (Heaven, Superimposition etc) is that the free versions don't allow "looping" the tests. (This applies to several other free utilities i know of as well.)

"The real problem is, is that it doesn't happen every time. I cannot reproduce it reliably. "

Since the OP has stated the issue is intermittent in nature, it's likely problematic to use something that can't be run at length. You would probably find, for instance, things appear to function normally during these short test runs, which would obviously be misleading. Won't hurt anything, but may not help with determining if there's something wrong with the GPU.

The only one I can think of that is free and still allows long run stress testing is FurMark, and some people think of it as dangerous (though it's probably OK if used properly).

As for using AfterBurner to manually control fans for testing, the problem there is that you'd be adding a source for controlling the fans that isn't usually present, thus it's not actually testing the normal state of the system.  Might work just fine; as above, it won't hurt anything...but might not help actually identifying whether there's a problem as indicated by the OP.

TBH, although many choose to do so for 'custom' controls, there is no need to use a third-party utility to control the fans adequately; card should be more than capable of that on it's own - which kinda takes us back to what the OP asked for: a utility for testing, as opposed to something to replace the built-in fan controls which appear to be malfunctioning.

To the OP: There might not be anything suitable to your purpose, and is free, and doesn't require Steam...you may need to reconsider your criteria there.  As the saying goes, "Beggars can't be choosers" 😉 

Best of luck.


Edited by kksnowbear
  • Like 1

Free professional advice: Do not rely upon any advice concerning computers from anyone who uses the terms "beast" or "rocking" to refer to computer hardware.  Just...don't.  You've been warned.

While we're at it, people should stop using the term "uplift" to convey "increase".  This is a technical endeavor, we're not in church or at the movies - and it's science, not drama.

Link to comment
Share on other sites

1 hour ago, kksnowbear said:

Of course; there are several free alternatives I'm aware of.

Problem with the Unigine benchmarks (Heaven, Superimposition etc) is that the free versions don't allow "looping" the tests. (This applies to several other free utilities i know of as well.)

 

Heaven runs always in a loop, whilst you are not benchmarking. Since you want to stress and not benchmark....

I use it all the time, when I want to heat saturate my water loop e.g.


Edited by Hiob
  • Like 1

"Muß ich denn jedes Mal, wenn ich sauge oder saugblase den Schlauchstecker in die Schlauchnut schieben?"

Link to comment
Share on other sites

27 minutes ago, Hiob said:

Heaven runs always in a loop, whilst you are not benchmarking.

Not according to Unigine, it doesn't:

image.png

EDIT:

Not sure why they say it doesn't loop, but I just checked a Basic installation and, as Hiob says above, it does appear to run (and load the the GPU) constantly.  If that's not considered 'looping', I wonder what they refer to in the chart above....


Edited by kksnowbear

Free professional advice: Do not rely upon any advice concerning computers from anyone who uses the terms "beast" or "rocking" to refer to computer hardware.  Just...don't.  You've been warned.

While we're at it, people should stop using the term "uplift" to convey "increase".  This is a technical endeavor, we're not in church or at the movies - and it's science, not drama.

Link to comment
Share on other sites

2 hours ago, kksnowbear said:

Not according to Unigine, it doesn't:

image.png

EDIT:

Not sure why they say it doesn't loop, but I just checked a Basic installation and, as Hiob says above, it does appear to run (and load the the GPU) constantly.  If that's not considered 'looping', I wonder what they refer to in the chart above....

 

because it doesn’t count frames (aka benchmarks) when it runs looping…

  • Like 1

"Muß ich denn jedes Mal, wenn ich sauge oder saugblase den Schlauchstecker in die Schlauchnut schieben?"

Link to comment
Share on other sites

15 minutes ago, Hiob said:

because it doesn’t count frames (aka benchmarks) when it runs looping…

Well, as stupid as that is...I have to acknowledge it behaves as it says 😄 😄  😄

Looking back, it seems I bought a license because the Superposition benchmark free version doesn't loop ("stress test").  Apparently the Heaven benchmark is different.

My apologies.

To the OP: It would seem this makes Heaven a good choice after all - sorry if my prior comments caused confusion.

 

  • Like 1

Free professional advice: Do not rely upon any advice concerning computers from anyone who uses the terms "beast" or "rocking" to refer to computer hardware.  Just...don't.  You've been warned.

While we're at it, people should stop using the term "uplift" to convey "increase".  This is a technical endeavor, we're not in church or at the movies - and it's science, not drama.

Link to comment
Share on other sites

Cinebench R24 does use both; CPU and GPU.

The tests run for 10min each, so it's kind of a little stress test per default benchmarking, no more "1 good run and hopefully not crash before" ( for the LN guys ).

 

https://www.maxon.net/en/downloads/cinebench-2024-downloads


Edited by BitMaster

Gigabyte Aorus X570S Master - Ryzen 5900X - Gskill 64GB 3200/CL14@3600/CL14 - Asus 1080ti EK-waterblock - 4x Samsung 980Pro 1TB - 1x Samsung 870 Evo 1TB - 1x SanDisc 120GB SSD - Heatkiller IV - MoRa3-360LT@9x120mm Noctua F12 - Corsair AXi-1200 - TiR5-Pro - Warthog Hotas - Saitek Combat Pedals - Asus PG278Q 27" QHD Gsync 144Hz - Corsair K70 RGB Pro - Win11 Pro/Linux - Phanteks Evolv-X 

Link to comment
Share on other sites

26 minutes ago, BitMaster said:

Cinebench R24 does use both; CPU and GPU.

The tests run for 10min each, so it's kind of a little stress test per default benchmarking, no more "1 good run and hopefully not crash before" ( for the LN guys ).

 

https://www.maxon.net/en/downloads/cinebench-2024-downloads

 

Interesting. I didn't notice that they released a new one. I'll have look, thanks.

"Muß ich denn jedes Mal, wenn ich sauge oder saugblase den Schlauchstecker in die Schlauchnut schieben?"

Link to comment
Share on other sites

vor 22 Stunden schrieb Hiob:

Interesting. I didn't notice that they released a new one. I'll have look, thanks.

"Muß ich denn jedes Mal, wenn ich sauge oder saugblase den Schlauchstecker in die Schlauchnut schieben?"

 

Haha, just saw your sig. I just watched that Episode on YT and had a laugh. Oh man, what a difference to most of today's comedians. 

"Hick's, noch 'nen Pahlgruber !?"

  • Like 1

Gigabyte Aorus X570S Master - Ryzen 5900X - Gskill 64GB 3200/CL14@3600/CL14 - Asus 1080ti EK-waterblock - 4x Samsung 980Pro 1TB - 1x Samsung 870 Evo 1TB - 1x SanDisc 120GB SSD - Heatkiller IV - MoRa3-360LT@9x120mm Noctua F12 - Corsair AXi-1200 - TiR5-Pro - Warthog Hotas - Saitek Combat Pedals - Asus PG278Q 27" QHD Gsync 144Hz - Corsair K70 RGB Pro - Win11 Pro/Linux - Phanteks Evolv-X 

Link to comment
Share on other sites

  • 2 months later...

So, for who wants to know how this story ended : had to send the GPU back for RMA. ASUS replaced the fans.

Managed to monitor the temps through the GPU TWEAK III program. It showed normal behaviour until it reached 70°C. Then the fans rpm-graph showed a saw-tooth profile ranging between 0 and max RPM, and after a while it would remain at 0 rpm permanently.

 

 

i7 6700k/GTX1070-8G/MSI-Z170A Gaming Pro Carbon/32GB DDR4 Kingston HyperX PREDATOR DDR4 3000MHZ Vengeance 1600/TM Warthog #6106/Samsung SB350_S27B350H/OCZ Agility3 SSD 128GB / Win10-64/TIR5



 

 

Link to comment
Share on other sites

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...