Frozen (crashed)view in VR

[chiliwili69]

I have been lately running at 4.9GHz and some moderate OC in the 1070, and I have been experimenting some frozen view in VR after 10 or 30 or 50 minutes (the time is random).

 

I thought it could be because heating the CPU/GPU too much and then I reduced to 4.8 GHz and monitorize temps, but the CPU temp is around 55 C when I play and GPU around 60 C. Look trend after half hour of play ion VR:

 

 

Another reason could be that I upgraded the Nvidia driver to 388, but I don´t know.

 

When it happens the view is frozen in VR but the sound of the plane continue until it crash the plane since I can not control. So, it should be something realted to SteamVR or similar (not IL2).

 

Any idea?

[dburne]

Running a 4790k at 4.9 GHz is really asking a lot from that chip.

 

I would try around 4.5, maybe 4.6 tops with some extra voltage.

[HunDread]

I think if the screen freezes and everything else goes on, it's the GPU. iI had this before and had to dial back GPU OC to solve it.

[chiliwili69]

Running a 4790k at 4.9 GHz is really asking a lot from that chip.

 

Why 4.9 GHz is a lot?

 

I am not an expert in OC but I thought that the limit in OC the CPU is determined by the temperatures you have under load. So, the voltage has to be increased to get stability until you hit the temperature constraint which is determined by how well the CPU evacuate the heat (thermal transfer coef, cooling system, etc).

 

I reported my OC results in this thread:

https://forum.il2sturmovik.com/topic/29881-overclocking-4790k-better-bos-performance/

 

It was also running well in 5.0 and 5.1 but I had to increase the voltage beyond 3.5, and I think is too much. So, 4.8 or 4.9 should be fine. Look this table for typical OC freq of the 4790K (the orange points are my CPU):

 

The 4790K has a base clock of 4.0 and turbo at 4.4.

The 7700K has a base clock of 4.2 and turbo at 4.5.

So, if the heat evacuation runs well, they should allow similar levels of OC. 

I think it is not CPU related since the game continue running.

Edited November 10, 2017 by chiliwili69

[A_radek]

I think if the screen freezes and everything else goes on, it's the GPU. iI had this before and had to dial back GPU OC to solve it.

I second that. Think I remember you oc:ing Gpu memory quite a bit. I myself seem to have gotten the wrong brand of Gpu memory so even a 200mhz increase gives errors.

 

Less likely, but perhaps have a look at total wattage during load? When both gpu and cpu are loaded it could be that your Psu is surrendering and ripples current after a while.

[chiliwili69]

Thanks for the tips.

The GPU OC that I was applying was +100MHz for GPU clock and +500MHz for VRAM. It was OK for 5 min of the Unigine Valley stress test (which is not VR) but 30 min of BOS VR could be more demanding.

I will try this weekend with less MHz for the VRAM.

[HunDread]

I started with something like + 200 GPU Clock and +400 VRAM. I had to dial back to around +50 GPU clock and +100 VRAM.

 

Anyway with the +200 GPU / +400 VRAM I gained about +1 FPS in the Balapan test. It doesn't really add anything it seems.

 

On the other hand CPU OC added a lot.

 

Forgot to add - I have a GTX 1070 and a 4790K

Edited November 10, 2017 by -[HRAF]Black_Sab

[=EXPEND=Tripwire]

I've started seeing similar issues in my Vive. For me the experience is like someone has ripped out the cables to the headset for 10 seconds. Then everything is back to normal.

 

Occurs once during a 2hr session, as a guess about 30mins in.

 

I have not updated drivers or changed any settings.

Edited November 10, 2017 by =TBAS=Tripwire

[Kraft]

Try the old drivers, I heard about the problems with 388 in IL-2. Maybe there is a problem with them.

[GER_GD]

Have also the same problem on my rig .

Fall creator update !?

Driver issue !?

I think it is not the oc cpu !

My i7 4790k is on stable 4.8 and the freezes starts with the fcu!

[dburne]

Thanks for the tips.

The GPU OC that I was applying was +100MHz for GPU clock and +500MHz for VRAM. It was OK for 5 min of the Unigine Valley stress test (which is not VR) but 30 min of BOS VR could be more demanding.

I will try this weekend with less MHz for the VRAM.

 

To rule it out you could always set the GPU back to default and test.

[dburne]

Why 4.9 GHz is a lot?

 

I am not an expert in OC but I thought that the limit in OC the CPU is determined by the temperatures you have under load. So, the voltage has to be increased to get stability until you hit the temperature constraint which is determined by how well the CPU evacuate the heat (thermal transfer coef, cooling system, etc).

 

 

 

Whilst the temperature is certainly a factor mainly on the protection of the CPU, it can absolutely not be stable and still have reasonable temperatures. Every chip is different in regards to stability and a certain voltage and overclocked settings.

 

When overclocking a CPU, I go up in small increments and then run Prime 95 small FFT's for a couple of hours at least, and make sure there are no errors and temps to not get too high at the amount of voltage I am putting into the CPU. If I get any errors I either lower the clockspeed or up the voltage, depending on how much voltage already giving it,

 

While that may well not be your issue, instability can certainly cause all sorts of weird problems. Same applies to the GPU. For the GPU I go up in small increments, testing with a few runs of Heaven Benchmark for stability as I go while monitoring GPU temps. The GPU is already being overclocked somewhat with Nvidia GPU Boost 3.0. With these Paschal cards memory overclocking can have diminishing returns.

 

A very good video guide for overclocking the Paschal cards can be seen here.

 

 

This guide was done using Precision XOC, however it is also can apply somewhat when using Afterburner.

Edited November 10, 2017 by dburne

[HunDread]

I've started seeing similar issues in my Vive. For me the experience is like someone has ripped out the cables to the headset for 10 seconds. Then everything is back to normal.

 

Occurs once during a 2hr session, as a guess about 30mins in.

 

I have not updated drivers or changed any settings.

 

I think this might be a different issue. From what I understand from the OP he has the screen frozen on a frame. This is what happened to me too and I fixed it by dialing back on the GPU OC.

The screen was freezing at some point. I still saw the picture (cockpit and environment) in the RIFT it was just not moving any more. Lost game movement and tracking movement too. And it was not coming back to normal. I had to restart my PC to recover.

[=EXPEND=Tripwire]

Yep I agree. I will move my issue to another thread which more closely matches my problem.

[chiliwili69]

When overclocking a CPU, I go up in small increments and then run Prime 95 small FFT's for a couple of hours at least, and make sure there are no errors and temps to not get too high at the amount of voltage I am putting into the CPU. If I get any errors I either lower the clockspeed or up the voltage, depending on how much voltage already giving it,   While that may well not be your issue, instability can certainly cause all sorts of weird problems. Same applies to the GPU. For the GPU I go up in small increments, testing with a few runs of Heaven Benchmark for stability as I go while monitoring GPU temps. The GPU is already being overclocked somewhat with Nvidia GPU Boost 3.0. With these Paschal cards memory overclocking can have diminishing returns.

 

Yeah, these were the steps I followed for the CPU and GPU overclock, but the GPU stress test I used was the Unigine Valley for 5 minutes. It seems that BOS in VR is more demanding.

 

 

Black_Sab' timestamp='1510319945' post='530532'] I think this might be a different issue. From what I understand from the OP he has the screen frozen on a frame. This is what happened to me too and I fixed it by dialing back on the GPU OC.

 

Problem solved! (so far).

I have just been flying 2 hours with no GPU OC with the CPU at 4.8 GHz and the problem didn´t appear.

 

So, from now on I will apply less OC to the GPU (or even nothing at all) since the fps gain is not noticeable at the game.

 

Thanks to all for your advice. 

Edited November 11, 2017 by chiliwili69

[=IL2AU=chappyj]

I've been having the same issue in the last 2 weeks.

It was fine before then

It may well be that removing the oc fixes it but it certainly became a problem after the latest patch so something changes that exposed the oc to instability in the software demands I guess

[katdog5]

Yeah, these were the steps I followed for the CPU and GPU overclock, but the GPU stress test I used was the Unigine Valley for 5 minutes. It seems that BOS in VR is more demanding.

 

 

 

Problem solved! (so far).

I have just been flying 2 hours with no GPU OC with the CPU at 4.8 GHz and the problem didn´t appear.

 

So, from now on I will apply less OC to the GPU (or even nothing at all) since the fps gain is not noticeable at the game.

 

Thanks to all for your advice. 

 

 

Yep that sounds right. I lock up too, pretty quickly   The threshold seems to be 110%, +200/+200 ish (or) 120% +200/+0 

 

although I haven't tested the 0 memory 120% yet..edit, yes I have. IL2 imm crash.  

 

see here: https://www.hardocp.com/article/2016/05/19/geforce_gtx_1080_founders_edition_overclocking_preview/

Edited December 21, 2017 by katdog5

[Lobstrosity]

I started with something like + 200 GPU Clock and +400 VRAM. I had to dial back to around +50 GPU clock and +100 VRAM.

 

Anyway with the +200 GPU / +400 VRAM I gained about +1 FPS in the Balapan test. It doesn't really add anything it seems.

 

On the other hand CPU OC added a lot.

 

Forgot to add - I have a GTX 1070 and a 4790K

 

I did the same thing for my 1080Ti. I was having the same issue in Elite Dangerous as well. I just dialed back the GPU overclock and that solved it.

[=EXPEND=Tripwire]

I know my issue was different to the OP, but in case anyone with a Vive comes across it - the fix was to disable power management. I haven't had the issue once in over a month now (where it was once a day).