2080 Ti - 1ms clock drop during FireStrike.

[sewerino26]

Hello.
I have a question. I am using Rtx 2080 Ti Waterforce and 9900K and Corsair 850RmX. ALL stock,no oc.
I run 3dmark firestrike 1920x1080. Monitor 1440P native.
Using Msi Afterburner not the newest,and riva tuner statistic.

Few days ago ,when i run Fire Strike GPU test 1 on loop, i saw a clock dip from 1980mhz to 1570mhz for 1ms. GPU LOAD not changed,it was 96%. Only gpu clock dip for a 1ms.
I checked in logs,and during that gpu power drop to 58%.


Log from MSI:

But then i look at gpuz and gpuz not report that,all was fine,no drop in clock and power.


I tried to replicate that. So i run in windowed mode Fire Strike and was messing in Windows, to force gpu power drops. It drops gpu power but clock was steady and not dropped,like before.

So why i cant reproduce that clock drop anymore? And why that happened? Should i worry?

 

[anxious_f0x]

Take it you didn’t get the answer you were looking for Here? https://forums.guru3d.com/threads/2080-ti-weird-gpu-power-spike-in-firestrike.428477/

 

[ljt]

I wouldn't be worrying about that, it might have been a glitch with Afterburner. It's such a short duration.

 

[sewerino26]

I ask because. I tried to replicate that. So i run in windowed mode Fire Strike and was messing in Windows, to force gpu power drops. It drops gpu power but clock was steady and not dropped,like before.

So why i cant reproduce that clock drop anymore? And why that happened? I dont ask about that. Yeah its my second topic there on GR, i dont wanna spam ,just to be sure. Eh

 

[ljt]

I wouldn't bother trying to replicate it. There's no need to analyse so much into one isolated incident that hasn't happened since. If everything is performing correctly then don't keep looking for something that isn't there. If you start getting corrupted images, artifacts or poor performance then start looking into it.

 

[koooowweeee]

Give it me and buy a new one. Problem solved

 

[Grim5]

Give it me and buy a new one. Problem solved

This it's obviously broken, I'll take it off your hands for a 10.

But seriously, OP, don't worry about it - especially if you can't replicate it.

 

[Gripen90]

Its Fire Strike acting up, it happens every now and then with the new updates for the benchmark. It is also a 6year old test that has a few oddities on newer hardware. I would use Time Spy or Port Royale instead.

 

[Kaapstad]

I would use Time Spy or Port Royale instead.

Pandora's box of questions incoming.

 

[sewerino26]

Its Fire Strike acting up, it happens every now and then with the new updates for the benchmark. It is also a 6year old test that has a few oddities on newer hardware. I would use Time Spy or Port Royale instead.

Yeah i know normal Fire Strike is old. But still good for checking stability overclocking on Rtx 2080/Ti. I dont oc my card of course.
So anyway that dip in clock was not caused by hardware? Even if i cant reproduce..?!

 

[Yaayuh!]

Firestrike is old and ****. I wouldn't bother with it any more. Just use newer benchmarks and games instead.

 

[jsbombjack]

Old benchmarks are really not useful
My pal ran 3DMark06 or similar against his 8700k/2080ti combo, and got a lower processor score than he did on his old athalon. Made me chuckle

 

[Kaapstad]

Yeah i know normal Fire Strike is old. But still good for checking stability overclocking on Rtx 2080/Ti. I dont oc my card of course.
So anyway that dip in clock was not caused by hardware? Even if i cant reproduce..?!

Firestrike standard is dreadful for checking anything because the fps are too high and bounce off of the CPU.

If you are going to use Firestrike at all use the Ultra version or better still as the other guys have said use something more modern.

What is even more comical is the game Civ 5 which is 9 years old is a far better test of stability than any synthetic benchmark as it really works the CPU and GPU over a long period of time and can generate a lot of heat in the process.

 

[sewerino26]

Firestrike standard is dreadful for checking anything because the fps are too high and bounce off of the CPU.

If you are going to use Firestrike at all use the Ultra version or better still as the other guys have said use something more modern.

What is even more comical is the game Civ 5 which is 9 years old is a far better test of stability than any synthetic benchmark as it really works the CPU and GPU over a long period of time and can generate a lot of heat in the process.

Ah ok thank you for explanation. We can close topic. No further replies needed.Thx guys

 

[sewerino26]

Hey guys. Just do not want to refresh all topic. I tested time spy gpu test 1.

I run TimeSpy looped gpu test 1. Clock during 40 minutes was 1920-1970mhz with random power usage. But one time it dropped to 1800mhz. Is this normal? Screens:

1920mhz:

And that 1800mhz for 1ms:

Of course its POWER LIMIT only in this test. So i guess its not faulty somewhere? In games i dont had any once such drops in clock.

 

[sewerino26]

Hey last question and we can close topic. So it was not related to faulty GPU?

 

[ljt]

Doesn't look like a faulty GPU, don't worry, just enjoy gaming on your system and forget about it.

 

[sewerino26]

Doesn't look like a faulty GPU, don't worry, just enjoy gaming on your system and forget about it.

Oki thank you for help and answers.Really apprecate it. We can close topic.