Incredible drop in frame rates today.

[spreckair]

I just tried to jump onto IL-2 today and it was not flyable.  My usual 70 to 90 FPS was dropping into the single digits, which has never happened before.  Since I only use my rig to play IL-2 and I haven't changed anything since my last flight last night, I wonder what is causing this.  Is anyone else having this issue?  Any suggestions are greatly appreciated.

[AEthelraedUnraed]

My first guess is your system jumped from dedicated to on-board GPU, for whatever reason. Check if your GPU works correctly. Also, if you use an NVIDIA GPU, make sure you've set the GPU to be the dedicated one for IL2 (there's probably similar settings for AMD cards).

[Riderocket]

 

36 minutes ago, AEthelraedUnraed said:

My first guess is your system jumped from dedicated to on-board GPU, for whatever reason. Check if your GPU works correctly. Also, if you use an NVIDIA GPU, make sure you've set the GPU to be the dedicated one for IL2 (there's probably similar settings for AMD cards).

I think this can only happen on a laptop, cause on desktop the dedicated and integrated GPUs have separate inputs for display.

 

Edit: OP what PC/Laptop do you have? Maybe it's something as simple as the thermal paste?

Edited April 7, 2022 by Riderocket

[spreckair]

Thank you for the replies.  I have a desktop PC.  I7-10700 CPU, 32 RAM, and an RTX3080 GPU.  Also, I am playing in VR--HP Reverb G2.  

 

Also, the CPU is an I7-10700KF, so it doesn't have integrated graphics.

Edited April 7, 2022 by spreckair

[Riderocket]

6 minutes ago, spreckair said:

Thank you for the replies.  I have a desktop PC.  I7-10700 CPU, 32 RAM, and an RTX3080 GPU.  Also, I am playing in VR--HP Reverb G2.

You could do a benchmark to monitor the CPU and GPU load, as well as temperature, if it's thermal throttling or not.

[spreckair]

26 minutes ago, Riderocket said:

You could do a benchmark to monitor the CPU and GPU load, as well as temperature, if it's thermal throttling or not.

Thank you Riderocket.  This is all new to me, so here goes.  Could you tell me what would be normal loads for CPU and GPU?  My GPU load was in the 90 percentile, but the CPU was running really low--teens and twenties.

[Riderocket]

30 minutes ago, spreckair said:

Thank you Riderocket.  This is all new to me, so here goes.  Could you tell me what would be normal loads for CPU and GPU?  My GPU load was in the 90 percentile, but the CPU was running really low--teens and twenties.

Hmm. It seems to be thermal throttling I'm guessing, Maybe you need to check the thermal paste and cooler and try again.

 

Wait sorry, I misunderstand, the GPU load was 90% or the benchmark results were 90%?

 

Usually the thing that is bottlenecking the system is at or close to 100% load and needs attention

Edited April 7, 2022 by Riderocket

[spreckair]

I think it is a thermal issue.  The CPU temps were spiking, which is sort of weird since the CPU was usage was not that high.

[Jaws2002]

Check your fans. Make sure they work.

[AEthelraedUnraed]

7 hours ago, Riderocket said:

Usually the thing that is bottlenecking the system is at or close to 100% load and needs attention

Benchmarks should usually bring your system to 100% load.

 

8 hours ago, spreckair said:

Thank you Riderocket.  This is all new to me, so here goes.  Could you tell me what would be normal loads for CPU and GPU?  My GPU load was in the 90 percentile, but the CPU was running really low--teens and twenties.

Here at the office (I work at an ICT desk) we usually use FurMark for stress tests, along with the included CPU burner (just open FurMark and click the "GPU stress test" and "CPU burner" buttons, then confirm for both). Basically, both your GPU and CPU usage should show 100%, unless it's throttling. To monitor your CPU/GPU temperatures, you could use a tool such as Speccy (there are many other similar programs). You'll likely see the CPU clock speed and "multiplier" decrease as the temperature increases - this is throttling going on.

 

Check the temperature, and how fast it rises (too sudden spikes can also be an indication of failing cooling). The absolute maximum temperature of your CPU/GPU depends on the exact type, but is usually around 100C or so. If it's really riding the edge, it might well be that there's a cooling problem; often it should start throttling at 10 or so degrees below that.

 

Could you report back?

[spreckair]

12 hours ago, AEthelraedUnraed said:

Could you report back?

I am running some tests now and will definitely report back when I know more.

[spreckair]

So my PC is sitting at 53C at idle, and is jumping to anywhere from the high 80C to 98C at just 25% usage.  I am leaning toward a cooling issue at this point.  Any thoughts are greatly appreciated.

Edited April 8, 2022 by spreckair

[AEthelraedUnraed]

14 minutes ago, spreckair said:

So my PC is sitting at 53C at idle, and is jumping to anywhere from the high 80C to 98C at just 25% usage.  I am leaning toward a cooling issue at this point.  Any thoughts are greatly appreciated.

Yeah, with those temperatures I'd say there's a cooling issue. Are your fans still running, and do they speed up properly if your PC gets hot? If so, try re-applying the cooling paste (a tube will cost €10 or so, there's tons of tutorials online about how to do it).

 

Since you mention the problem started pretty sudden, I don't think dust is the issue here, but it's a good idea to periodically blow out all dust regardless.

[spreckair]

1 hour ago, AEthelraedUnraed said:

Yeah, with those temperatures I'd say there's a cooling issue. Are your fans still running, and do they speed up properly if your PC gets hot? If so, try re-applying the cooling paste (a tube will cost €10 or so, there's tons of tutorials online about how to do it).

 

Since you mention the problem started pretty sudden, I don't think dust is the issue here, but it's a good idea to periodically blow out all dust regardless.

Thank you for the assist.  I just got a tube of cooling paste so fingers crossed.  Otherwise, it will be another IT project for the novice--me.

[Jaws2002]

Before you reapply the paste, physically check all the fan connections, make sure all the fans are properly connected into the motherboard, including the CPU fan.  Make sure they work. Check their RPM with some monitoring software, like HWiNFO.

Only after you are sure the fans work properly, check the paste.

I did find fans disconnected from the motherboard after cleaning and I also found dead fans after restart.

[Strewth]

37 minutes ago, spreckair said:

Thank you for the assist.  I just got a tube of cooling paste so fingers crossed.  Otherwise, it will be another IT project for the novice--me.

If you have done the paste thing successfully before, then great.
If you are not experienced in paste application, then I would recommend some research.
Not trying to lecture, but:
Clean clean clean.
Not too much and not too little.
There are different patterns with different results.
Cheers and best of luck.

 

[spreckair]

5 hours ago, Strewth said:

Cheers and best of luck.

 

Thank you Strewth; I have been doing research.  I can't screw this up if I want to fly.

[Strewth]

On 4/8/2022 at 7:34 AM, spreckair said:

Thank you for the replies.  I have a desktop PC.  I7-10700 CPU, 32 RAM, and an RTX3080 GPU.  Also, I am playing in VR--HP Reverb G2.  

 

Also, the CPU is an I7-10700KF, so it doesn't have integrated graphics.

Also. What cooling are you using?
I originally used a Ryzen 7 3800XT, but upgraded to a Ryzen 9 5950X. With both setups I use a Corsair H100X water cooling system and at the time it cost me the equivalent of around US$80-00. Liquid cooling is a great investment on a lot of the modern chips. And most come with universal mounts, as well as a thermal paste pre-coated pad.
If you haven't already got liquid cooling, it may be worth a look.

Cheers

[spreckair]

Sorry all, I just had a gall bladder removed and I am a bit laid up.  I will get back to this as soon as I am able.  Thank you for your attention.

[spreckair]

 

On 4/9/2022 at 2:05 AM, Strewth said:

If you haven't already got liquid cooling, it may be worth a look.

So, to continue this conversation, I re-pasted the 120mm AIO CPU cooler but it didn't help the issue.  After researching, it looks like I need a new case and larger cooler for my rig, which is a pre-built HP Omen 30L and is known for having cooling issues.  The new case (Cooler Master MasterBox MB 311L) and EK AIO Basic 240 mm cooler are ordered and on the way.  This is yet another new unexpected adventure for me, so wish me luck.  

[spreckair]

So, to round this out, I ended up just building a new rig.  New parts:  Fractal case, I-7 11700K CPU, Crucial 32 G RAM, MSI motherboard, 240MM AIO cooler; old parts:  500GB SSD, 750W power supply, RTX 3080.  I am back in the cockpit and everything is now running smooth and cool...for now.  Fingers crossed.