Jump to content
chiliwili69

Measuring rig performance: Common Baseline (for IL-2 v3.010)

Recommended Posts

I got a lower average FPS with the new 3.006 patch. Is anyone else seeing this? My Spitfire bomber run went from ~87.5 fps average down to ~83 fps average.

 

On the plus side, I've just tried OpenComposite (the alternative to SteamVR) and it seems to give me back some fps as well as improving on the 0.1% lows by quite a lot. If anyone else wants to try OpenComposite, do this:

5 hours ago, chiliwili69 said:

Thanks for the results, but in order to be sure about the result, you can run the test 2 or 3 times in order to see if you obtain almost identical results. So, I can write your result to the table.

 

Chili I think you should re-run your benchmark with the new patch. I'm showing reduced FPS on my system with the new IL-2 patch as of today. If that's a general trend (and not just my machine) that would unfortunately invalidate a lot of those older benchmarks... 😞

  • Upvote 2

Share this post


Link to post
Share on other sites
53 minutes ago, Alonzo said:

I got a lower average FPS with the new 3.006 patch. Is anyone else seeing this? My Spitfire bomber run went from ~87.5 fps average down to ~83 fps average.

 

On the plus side, I've just tried OpenComposite (the alternative to SteamVR) and it seems to give me back some fps as well as improving on the 0.1% lows by quite a lot. If anyone else wants to try OpenComposite, do this:

 

Chili I think you should re-run your benchmark with the new patch. I'm showing reduced FPS on my system with the new IL-2 patch as of today. If that's a general trend (and not just my machine) that would unfortunately invalidate a lot of those older benchmarks... 😞

How much did you gain? I thought my game felt smoother with the new update I didn't bench anything though. Also do you need oculus home running to use this?

Edited by 15th_JonRedcorn

Share this post


Link to post
Share on other sites
53 minutes ago, Alonzo said:

On the plus side, I've just tried OpenComposite (the alternative to SteamVR) and it seems to give me back some fps as well as improving on the 0.1% lows by quite a lot.

 

Having played around a bit, I'm finding a bit of head tracking judder when using OpenComposite -- anyone tried it, seeing the same thing? The benchmark numbers look good but I'm not sure the in-game experience matches the (apparently good) numbers.

1 minute ago, 15th_JonRedcorn said:

How much did you gain? I thought my game felt smoother with the new update I didn't bench anything though. Also do you need oculus home running to use this?

 

I lost about 4 fps on average with the new patch, and gained back about 3 fps using OpenComposite. But the true test is in the playing of it. You don't need Oculus Home to be running -- just quit IL2 and SteamVR, rename the DLL, copy in the new OpenComposite DLL and create the .ini file. Then start IL2 as usual and you will see that SteamVR doesn't start this time, because OpenComposite is going straight to the Oculus drivers.

Share this post


Link to post
Share on other sites
8 minutes ago, Alonzo said:

 

Having played around a bit, I'm finding a bit of head tracking judder when using OpenComposite -- anyone tried it, seeing the same thing? The benchmark numbers look good but I'm not sure the in-game experience matches the (apparently good) numbers.

 

I lost about 4 fps on average with the new patch, and gained back about 3 fps using OpenComposite. But the true test is in the playing of it. You don't need Oculus Home to be running -- just quit IL2 and SteamVR, rename the DLL, copy in the new OpenComposite DLL and create the .ini file. Then start IL2 as usual and you will see that SteamVR doesn't start this time, because OpenComposite is going straight to the Oculus drivers.

I just tried it, felt like it was running much better than using steamvr I dig it. Not only that but it fixed lufentsa's 3d zoom mod that wasn't working while using the new update and steamvr. I didn't notice any tracking issues. Heck I had firefox running while I was playing too.

Share this post


Link to post
Share on other sites
23 minutes ago, 15th_JonRedcorn said:

I just tried it, felt like it was running much better than using steamvr I dig it. Not only that but it fixed lufentsa's 3d zoom mod that wasn't working while using the new update and steamvr. I didn't notice any tracking issues. Heck I had firefox running while I was playing too.

 

How many sensors do you have? I have 2 sensors with the Touch controllers. Am wondering if there's something happening with my tracking.

Share this post


Link to post
Share on other sites
9 hours ago, Alonzo said:

On the plus side, I've just tried OpenComposite (the alternative to SteamVR) and it seems to give me back some fps as well as improving on the 0.1% lows by quite a lot. If anyone else wants to try OpenComposite, do this:

 

Sounds promising... Does this work with non-Steam version of IL-2?

 

Thanks

Share this post


Link to post
Share on other sites
1 hour ago, -[HRAF]Black_Sab said:

Sounds promising... Does this work with non-Steam version of IL-2?

 

It ought to, but I only have the Steam version so I can't test that (or at least I have only been able to get IL-2 to start from the Steam launcher for me, but that might just be me being inept).

Share this post


Link to post
Share on other sites
16 hours ago, Alonzo said:

 

How many sensors do you have? I have 2 sensors with the Touch controllers. Am wondering if there's something happening with my tracking.

I use two sensors, make sure they aren't too close together and also make sure you have no monitoring software polling the graphics card. That alleviated most of my tracking problems, was having a ton of tracking issues like 2 months ago.

Share this post


Link to post
Share on other sites
19 hours ago, Alonzo said:

I got a lower average FPS with the new 3.006 patch. Is anyone else seeing this? My Spitfire bomber run went from ~87.5 fps average down to ~83 fps average.

 

On the plus side, I've just tried OpenComposite (the alternative to SteamVR) and it seems to give me back some fps as well as improving on the 0.1% lows by quite a lot. If anyone else wants to try OpenComposite, do this:

 

Chili I think you should re-run your benchmark with the new patch. I'm showing reduced FPS on my system with the new IL-2 patch as of today. If that's a general trend (and not just my machine) that would unfortunately invalidate a lot of those older benchmarks... 😞

The URL is not working

7 hours ago, Alonzo said:

 

It ought to, but I only have the Steam version so I can't test that (or at least I have only been able to get IL-2 to start from the Steam launcher for me, but that might just be me being inept).

You should be able to install it anyways if you have activated the keys:

http://cdn.il2sturmovik.net/x64/IL2_setup_BoM.exe

Share this post


Link to post
Share on other sites

I've started a new topic concerning OpenComposite. For me at least, it has actually made a very noticeable difference.

To address a couple points in this thread, first, it works perfectly with the non-steam version (and if you do the Oculus Home run-as-admin trick, it won't even launch Oculus - saving you piles of memory from not needing Steam running, and also not needing Oculus Home running). Second, I did some more testing with the newest DLL, and the head track judder is no longer noticeable. The DLL is updated very often, as in several times per day, so get the newest one.

  • Upvote 1

Share this post


Link to post
Share on other sites
7 hours ago, E69_Qpassa said:

The URL is not working

 

Works fine for me, I just clicked it. It immediately attempts to do a download - do you have any browser download stuff installed?

  • Thanks 1

Share this post


Link to post
Share on other sites
21 hours ago, Alonzo said:

 

It ought to, but I only have the Steam version so I can't test that (or at least I have only been able to get IL-2 to start from the Steam launcher for me, but that might just be me being inept).

Yes I tried and works fine with non-steam version. 

Share this post


Link to post
Share on other sites

Does anyone have problems playing longer then an hour at higher clock speeds .  8086k 4.8ghz I have to back out after each online sortie or my system crashes.  The weird thing is my temps are in the mid 50's low 60's and everything is stable only running 1.21vcore, 44ring,adaptive -.04. If I push is to an hour and a half online my computer locks up.  

 

Also anyone tried benchmarking with enhanced turbo boost turned on?  I enabled it last night just messing around and it seemed like my computer was going to grenade.  

 

Another thing is has anyone tried uninstalling the nvidia physx software?  It might just be me but it feels like Im getting a clearer image with less shimmer with that uninstalled leaving just the gpu driver. Driver is 388.00, no performance gain measured. 

Edited by JV44_Megla

Share this post


Link to post
Share on other sites
18 hours ago, Alonzo said:

 

Works fine for me, I just clicked it. It immediately attempts to do a download - do you have any browser download stuff installed?

now its working, thanks

Share this post


Link to post
Share on other sites
7 hours ago, JV44_Megla said:

Does anyone have problems playing longer then an hour at higher clock speeds .  8086k 4.8ghz I have to back out after each online sortie or my system crashes.  The weird thing is my temps are in the mid 50's low 60's and everything is stable only running 1.21vcore, 44ring,adaptive -.04. If I push is to an hour and a half online my computer locks up.  

 

Also anyone tried benchmarking with enhanced turbo boost turned on?  I enabled it last night just messing around and it seemed like my computer was going to grenade.  

 

Another thing is has anyone tried uninstalling the nvidia physx software?  It might just be me but it feels like Im getting a clearer image with less shimmer with that uninstalled leaving just the gpu driver. Driver is 388.00, no performance gain measured. 

 
The workflow of overclocking is:

1: Get Prime95 or similar. (it may be outdated by now? been a while for me), monitor or log temperatures.
2: Increase Mhz

3: Test with Prime95

4: Repeat 2 and 3 until it resets

5: If it reset whilst not hitting temp limits then either lower the mhz and leave it  there OR increase voltage 

(End if stable and happy with OC, however if you increased voltage continue)

6: Rerun Prime95 with increased voltage until its stable (pay attention to temps!)

7: Once stable with higher voltage and temps fine, start again at step 2.
8: Repeat until temps near limits

 

If your temps are still low during a stress test (not a game, do a real test) then firstly double check you're getting the right measurements (or else you really will cause damage) or start at step 5.


Nvidia physx driver shouldn't be in use for IL2, and so there should be no difference on vs off.

Share this post


Link to post
Share on other sites
7 hours ago, JV44_Megla said:

Does anyone have problems playing longer then an hour at higher clock speeds .  8086k 4.8ghz I have to back out after each online sortie or my system crashes.  The weird thing is my temps are in the mid 50's low 60's and everything is stable only running 1.21vcore, 44ring,adaptive -.04. If I push is to an hour and a half online my computer locks up.  


I was helping someone overclock their rig today. It's important to do a Prime95 test with CPU-Z and CoreTemp running, to check your actual CPU frequency. 

 

On the rig we OC'd today, the CPU would start throttling after a little while. It turned out to be the CPU power limits, which limit CPU current and boost clock after a while. We did these steps:

  • Increase the CPU power limits

    • Switch OFF Intel Adaptive Thermal Monitor (if your motherboard has this setting)

    • Set Long Duration Power Limit, Long Duration Maintained, Short Duration Power Limit and CPU Current Limit to maximum values. Just type 99999 into the text box and the motherboard will set it to like 4096 or 256 depending on the value.

If you're interested in the rest of our overclocking tips they are here: https://docs.google.com/document/d/1CBd9ubU78ZScBhEDry7I_p3_irptAVTr09gTqPIf0Ls/edit#

Share this post


Link to post
Share on other sites
2 hours ago, peregrine7 said:

 
The workflow of overclocking is:

1: Get Prime95 or similar. (it may be outdated by now? been a while for me), monitor or log temperatures.
2: Increase Mhz

3: Test with Prime95

4: Repeat 2 and 3 until it resets

5: If it reset whilst not hitting temp limits then either lower the mhz and leave it  there OR increase voltage 

(End if stable and happy with OC, however if you increased voltage continue)

6: Rerun Prime95 with increased voltage until its stable (pay attention to temps!)

7: Once stable with higher voltage and temps fine, start again at step 2.
8: Repeat until temps near limits

 

If your temps are still low during a stress test (not a game, do a real test) then firstly double check you're getting the right measurements (or else you really will cause damage) or start at step 5.


Nvidia physx driver shouldn't be in use for IL2, and so there should be no difference on vs off.

 

Peregrine thanks for the tips. 

 

I do use prime95 to stress test and hwmonitor to track.  temps at 100% optimization were stable at 73c after 30min..  Brought down the voltage from 127 to 125 and switched from adaptive to disabled. Ran again and everything stat's wise looks the same except a decrease voltage in temp down to 70c. Played for 2 hours tonight so far no issues. Pretty sure the adaptive voltage was my instability.  

 

Agreed on the physx. I knew it didnt have an affect on il2 but rendering seems soother with it uninstalled all together. Might all be in my head though.

2 hours ago, Alonzo said:


I was helping someone overclock their rig today. It's important to do a Prime95 test with CPU-Z and CoreTemp running, to check your actual CPU frequency. 

 

On the rig we OC'd today, the CPU would start throttling after a little while. It turned out to be the CPU power limits, which limit CPU current and boost clock after a while. We did these steps:

 

  • Increase the CPU power limits

    • Switch OFF Intel Adaptive Thermal Monitor (if your motherboard has this setting)

    • Set Long Duration Power Limit, Long Duration Maintained, Short Duration Power Limit and CPU Current Limit to maximum values. Just type 99999 into the text box and the motherboard will set it to like 4096 or 256 depending on the value.

If you're interested in the rest of our overclocking tips they are here: https://docs.google.com/document/d/1CBd9ubU78ZScBhEDry7I_p3_irptAVTr09gTqPIf0Ls/edit#

 

Alonzo thank for the help as well and thanks for the download. I surely will give it a read.  Looks like Adaptive might have been my issue.   

 

Also I tried to run prime95 with enhanced turbo boost/mse  on.   Crashed instantly.. wouldnt recommend anyone even bother enabling that setting if have.  

 

 

 

 

Share this post


Link to post
Share on other sites
3 hours ago, Alonzo said:

Increase the CPU power limits

  • Switch OFF Intel Adaptive Thermal Monitor (if your motherboard has this setting)

  • Set Long Duration Power Limit, Long Duration Maintained, Short Duration Power Limit and CPU Current Limit to maximum values. Just type 99999 into the text box and the motherboard will set it to like 4096 or 256 depending on the value.

 

Yes, this is important as far as temps are OK.

These items are included as well in the BIOS settings that I was detailing in this post:

https://forum.il2sturmovik.com/topic/34107-items-to-review-if-you-have-low-performance-in-il-2-vr-test/

 

at the bottom of the post you will find the BIOS settings I am using (where these settings are highlighted):

https://docs.google.com/spreadsheets/d/1zrEWtXGD4d1KMG3v9bGLeyITKJOAzA2Q91ghgCs0NxY

 

3 hours ago, peregrine7 said:

Test with Prime95

 

Yes, Prime95 is quite valid and simple for CPU stress test. After all is just crushing numbers.

But it is important what test of Prime95 you run. I was talking about this before:

 

https://forum.il2sturmovik.com/topic/29322-measuring-rig-performance-common-baseline-for-il-2-v3005/?do=findComment&comment=514455

 

and also talking about what Prime95 does and what test we should run:

https://forum.il2sturmovik.com/topic/29322-measuring-rig-performance-common-baseline-for-il-2-v3005/?do=findComment&comment=549855

 

  • Like 1

Share this post


Link to post
Share on other sites
On 9/27/2018 at 3:27 AM, Alonzo said:

I got a lower average FPS with the new 3.006 patch. Is anyone else seeing this? My Spitfire bomber run went from ~87.5 fps average down to ~83 fps average

 

Yes, I also observed a lower fps with the new 3.006. It is about 5 or 6 fps less!!.

These are my test:

(The only note is that with the 3.006 I was running also MSI afterburner, but I am sure this impact is very small, maybe just 1 fps)

 

3.005 version Becnhmark settings
2018-09-14 18:32:18 - Il-2
Frames: 7336 - Time: 120000ms - Avg: 61.133 - Min: 43 - Max: 91

 

3.006 version Benchmark settings
2018-09-29 07:18:44 - Il-2
Frames: 6711 - Time: 120000ms - Avg: 55.925 - Min: 43 - Max: 91
Frames: 6663 - Time: 120000ms - Avg: 55.525 - Min: 43 - Max: 91
Frames: 6549 - Time: 120000ms - Avg: 54.575 - Min: 43 - Max: 91
Frames: 6536 - Time: 120000ms - Avg: 54.467 - Min: 42 - Max: 91
Frames: 6314 - Time: 120000ms - Avg: 52.617 - Min: 42 - Max: 83 Discarded
Frames avg: 6615
Avg fps: 55.123
 

366184700_version3.005-3_006.thumb.jpg.6ec105475663b3916189b763d476a663.jpg

I have done some test to determine influence of GRASS and SHADOWS in the benchmark.

https://forum.il2sturmovik.com/topic/40047-fps-impact-of-grass-quality-and-shadows/

 

 

Share this post


Link to post
Share on other sites

I bought a 2080 ti but I have a problem.

 

Months ago with my 1080 ti, before the Steam VR update for supersampling, I would use Oculus Tray Tool. I would set IL-2 to 1.5 ss. I would see frame rates fluctuate 50 and 80 fps. 

 

I don't know if it was the Steam VR update or some IL-2 update but now I'm stuck at 45 fps average. Here are my test results with 150% supersampling following the guidelines

 

2018-09-30 22:12:38 - Il-2
Frames: 5340 - Time: 120000ms - Avg: 44.500 - Min: 34 - Max: 64

 

Fallout 4 VR and Steam VR both perform badly. Not sure what's going on. 

Edited by BlackMambaMowTin

Share this post


Link to post
Share on other sites
9 hours ago, BlackMambaMowTin said:

I bought a 2080 ti but I have a problem

 

Thanks you for your results, but please, describe your remaining system (CPU, overclock, etc) as required in the instructions of first post (Giving Passmark numbers).

It will be good to have the first results of the benchmark with a 2080Ti.

 

You can also take a look of this post to review your low results:

https://forum.il2sturmovik.com/topic/34107-items-to-review-if-you-have-low-performance-in-il-2-vr-test/

 

Share this post


Link to post
Share on other sites
9 hours ago, BlackMambaMowTin said:

I bought a 2080 ti but I have a problem.

 

Months ago with my 1080 ti, before the Steam VR update for supersampling, I would use Oculus Tray Tool. I would set IL-2 to 1.5 ss. I would see frame rates fluctuate 50 and 80 fps. 

 

I don't know if it was the Steam VR update or some IL-2 update but now I'm stuck at 45 fps average. Here are my test results with 150% supersampling following the guidelines

 

2018-09-30 22:12:38 - Il-2
Frames: 5340 - Time: 120000ms - Avg: 44.500 - Min: 34 - Max: 64

 

Fallout 4 VR and Steam VR both perform badly. Not sure what's going on. 

Did you just swap the GPUs and then install the new driver? Or did you use the DDU tool? Often enough, there are leftovers from the old GPU driver and this can have an impact. I find using DDU is the best thing next to complete new install of the rig.

Share this post


Link to post
Share on other sites
6 hours ago, chiliwili69 said:

 

Thanks you for your results, but please, describe your remaining system (CPU, overclock, etc) as required in the instructions of first post (Giving Passmark numbers).

It will be good to have the first results of the benchmark with a 2080Ti.

 

You can also take a look of this post to review your low results:

https://forum.il2sturmovik.com/topic/34107-items-to-review-if-you-have-low-performance-in-il-2-vr-test/

 

 

My CPU is an i7 6700K @ 4.00GHz 16GB dual channel don't remember specifics.

 

I'll do a cpuz and give the full numbers later.

 

@ZachariasX

 

I didn't use the DDU tool but I'll give it a try. I'll also reinstall Steam VR since I think the problem is there. 

 

Question: If you minimize all the settings in IL-2, do you get a stable 90 fps? 

Share this post


Link to post
Share on other sites
2 hours ago, BlackMambaMowTin said:

Question: If you minimize all the settings in IL-2, do you get a stable 90 fps?

 

Yes, if you just put pre-set to LOW and leave all other settings to low or medium (and mirrors off), you will most likely have 90fps all time.

Then you can try to increase one by one until your performance is affected.

When you run the becnhmark, you can also run MSI Afterburner (it is a free tool which is thought for GPU overclock, but I used it just for trending the GPU load, GPU temp, CPU temp and CPU frequency. In options you can configure it). So, with MSI afterbuner you will see that your GPU load should be always at 50% or below for 150% SteamVR supersampling.

 

Also, you will benefit a lot if you overclock your CPU to 4.6 o perhaps more. Doing overclock is pretty easy. You can find the procedure here:

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

 

Having a 2080Ti and not doing overclocking for IL-2 is a tragedy.

Share this post


Link to post
Share on other sites
23 hours ago, BlackMambaMowTin said:

I bought a 2080 ti but I have a problem.

 

Months ago with my 1080 ti, before the Steam VR update for supersampling, I would use Oculus Tray Tool. I would set IL-2 to 1.5 ss. I would see frame rates fluctuate 50 and 80 fps. 

 

I don't know if it was the Steam VR update or some IL-2 update but now I'm stuck at 45 fps average. Here are my test results with 150% supersampling following the guidelines

 

SteamVR sucks at setting proper supersampling and often conflicts with itself and with the Oculus tools. Try installing Oculus Tray Tool and setting "Visible HUD" to "Pixel density". On my rig this showed that even though I *thought* I was using a low amount of supersampling, I was actually using quite a lot more. This taxes the GPU and may be a reason you're stuck at 45 fps. I would advise forcing SteamVR to 100% supersampling (both global and for IL2) and using *just* Oculus Tray Tool to set a global supersample.

Share this post


Link to post
Share on other sites
7 hours ago, chiliwili69 said:

 

Yes, if you just put pre-set to LOW and leave all other settings to low or medium (and mirrors off), you will most likely have 90fps all time.

Then you can try to increase one by one until your performance is affected.

When you run the becnhmark, you can also run MSI Afterburner (it is a free tool which is thought for GPU overclock, but I used it just for trending the GPU load, GPU temp, CPU temp and CPU frequency. In options you can configure it). So, with MSI afterbuner you will see that your GPU load should be always at 50% or below for 150% SteamVR supersampling.

 

Also, you will benefit a lot if you overclock your CPU to 4.6 o perhaps more. Doing overclock is pretty easy. You can find the procedure here:

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

 

Having a 2080Ti and not doing overclocking for IL-2 is a tragedy.

 

I set Global and App supersampling to 100%. All reprojection whatever unchecked. IL-2 setting on LOW defaults. ASW off. Still terrible framerate. At 45 fps most of time and dipping to 30 fps sometimes. 

 

I also reinstalled SteamVR and turned off OTT. 

Share this post


Link to post
Share on other sites
4 hours ago, Alonzo said:

SteamVR sucks at setting proper supersampling and often conflicts with itself and with the Oculus tools. Try installing Oculus Tray Tool and setting "Visible HUD" to "Pixel density". On my rig this showed that even though I *thought* I was using a low amount of supersampling, I was actually using quite a lot more. This taxes the GPU and may be a reason you're stuck at 45 fps. I would advise forcing SteamVR to 100% supersampling (both global and for IL2) and using *just* Oculus Tray Tool to set a global supersample.

The problem is to use both OTT and SteamVR.

I have been always using SteamVR to set SS and it has worked fine so far. To check the actual resolution that your display is using, instead of using OTT (which is not an official Oculus software) I was using Oculus Mirror with Oculus Debug Tool (both native software of Oculus) with "Visible HUD" enabled to check the actual rendered resolution to the Rift.

This is explained in this post:

https://forum.il2sturmovik.com/topic/30436-pixel-density-supersampling-at-steamvr-oculustraytool-and-oculusdebugtool/

 

Try to uninstall OTT and all your problems will disappear.

Share this post


Link to post
Share on other sites
21 hours ago, chiliwili69 said:

The problem is to use both OTT and SteamVR.

I have been always using SteamVR to set SS and it has worked fine so far. To check the actual resolution that your display is using, instead of using OTT (which is not an official Oculus software) I was using Oculus Mirror with Oculus Debug Tool (both native software of Oculus) with "Visible HUD" enabled to check the actual rendered resolution to the Rift.

This is explained in this post:

https://forum.il2sturmovik.com/topic/30436-pixel-density-supersampling-at-steamvr-oculustraytool-and-oculusdebugtool/

 

Try to uninstall OTT and all your problems will disappear.

 

I tried all of that. The render resolution seems fine. 

 

The odd thing is my GPU utilization is around 40% no matter what (also low CPU utilization). I have it set a performance globally in Nvidia control but still. 

 

I'm ready to rip my hear out with frustration. Maybe this weekend I should just reinstall windows. 

 

Again, no matter how low I set things my frame rate is stuck.

 

P.S. PassdMark CPU Mark 10971.4 (88th percentile) i7 6700K @ 4.0 GHz

Overall 5309.2 (94th percentile)

Edited by BlackMambaMowTin

Share this post


Link to post
Share on other sites
56 minutes ago, BlackMambaMowTin said:

P.S. PassdMark CPU Mark 10971.4 (88th percentile) i7 6700K @ 4.0 GHz

 

If your CPU is indeed only running at 4.0 GHz, it is far too slow, i.e., the best CPUs for Il-2 are some 25 percent faster. The PassMark CPU score is slightly worse than a typical 6700K. It's no surprise that you're getting poor performance.

 

Unless you have strangely inadequate equipment, getting 4.6 GHz or more is trivial.

 

56 minutes ago, BlackMambaMowTin said:

The odd thing is my GPU utilization is around 40% no matter what (also low CPU utilization).

 

Low GPU utilization is indicative of a CPU-bound situation. Low CPU utilization does not mean much.

 

56 minutes ago, BlackMambaMowTin said:

Maybe this weekend I should just reinstall windows. 

 

That's a very sensible thing to do from time to time.

Edited by Mitthrawnuruodo

Share this post


Link to post
Share on other sites

 

1 hour ago, BlackMambaMowTin said:

Again, no matter how low I set things my frame rate is stuck.

 

Definitely your cpu. If you disable asw through oculus debug tool (inside: program files/oculus/support/diagnostics) you should see your frame rate briefly jump above 45fps. At least at low settings, in an empty quick mission and when up high looking at sky only. The only alternative culptrit here would be that something else also is strangely wrong or corrupted, but once you find and fix that your cpu will still hold you back to 45fps.

 

Like the others said a 6700 is no slouch and can reach 4.6 quite easily. Do try to run cpu-z as you have il2 running as that, in real time, will show you what clockspeed your at. If for instance there is a an Avx offset set in your bios your cpu could be throttling down to 3.8ghz when running il2.

 

56 minutes ago, Mitthrawnuruodo said:

Low GPU utilization is indicative of a CPU-bound situation. Low CPU utilization does not mean much.

 

What he said.

Share this post


Link to post
Share on other sites
2 hours ago, BlackMambaMowTin said:

PassdMark CPU Mark 10971.4 (88th percentile) i7 6700K @ 4.0 GHz

 

For people running the Passmark benchmark I ask them just two values, as indicated in the instructions:

CPU Mark (this number sumarize how good is your CPU for all kind of Multithread tasks, it is not very important for IL-2)

CPU Single-Threaded (this number tells you how good is your CPU for single-thread tasks, like IL-2, which is very dependent of one thread, this is the number that matters)

post-18865-0-47048300-1504118846.thumb.jpg.11971ba78c8fc65e13fb0da6dfd4774d.jpg

2 hours ago, BlackMambaMowTin said:

also low CPU utilization

 

Don´t trust on the CPU utilization numbers that the monitoring applications delivers to you. They are fake. Even if you only look to the utilization of one core. This is because the thread are jumping from core to core, so the average is low, but it doesn´t mean your CPU is not the guilty guy.

2 hours ago, BlackMambaMowTin said:

GPU utilization is around 40%

 

In that case the problem is somewhere else. You can try to:

 

Report in this thread:

Motherboard brand and model

Speed fo your 16GB RAM, how many modules (is it 2x8Gb or 4x4Gb or 1x16Gb?)

CPU cooling system

 

Then, try to go through all these items:

https://forum.il2sturmovik.com/topic/34107-items-to-review-if-you-have-low-performance-in-il-2-vr-test/

 

Then, do overcloking. It is an easy thing. (I did it and I am not a guru)

Share this post


Link to post
Share on other sites
On 10/2/2018 at 3:16 AM, chiliwili69 said:

The problem is to use both OTT and SteamVR.

 

Try to uninstall OTT and all your problems will disappear.

 

...except that SteamVR will double-apply SS settings if you restart IL-2 (rather than shut down IL-2, shut down SteamVR, and restart).

 

Honestly it doesn't matter which tool people use, as long as they are checking the actual Pixel Density using the HUD (or using your method with Oculus Mirror).

10 hours ago, BlackMambaMowTin said:

I tried all of that. The render resolution seems fine. 

 

Do you mean "I switched on the Pixel Density HUD and it told me Pixel Density 1.1" or something else? "Seems fine" is not the same as actually looking in the HUD.

Share this post


Link to post
Share on other sites

Hello there !

 

I don't want to ruin the party here, especially since I'm much less experienced about CPU/RAM overclocking, but according to BlackMamba's first post he was getting better performances with his previous 1080Ti :

 

On 10/1/2018 at 4:27 AM, BlackMambaMowTin said:

Months ago with my 1080 ti, before the Steam VR update for supersampling, I would use Oculus Tray Tool. I would set IL-2 to 1.5 ss. I would see frame rates fluctuate 50 and 80 fps. 

 

Everything else considered equal, his 2080Ti should perform at least as good right ?

 

For the clarity of this (already too long) thread, I'd suggest to move the current discussion to a dedicated 2080Ti performance troubleshooting one. 

 

Thanks again for your precious contributions.

 

Share this post


Link to post
Share on other sites

Following up in our discussion in pimax topic about 9900k I can say that Intel smartcache allows a single core to use the whole L3 cache.

The question is how much is performance affected by the L3 cache size. If the frequencies are equal then cache change from 8700 to 9900 is nearly the same as from 8600 to 8700. So we need to compare 8600 and 8700 at the same CPU and RAM frequency.

Share this post


Link to post
Share on other sites

Well, we have almost what you ask for your test. the 8700K and 8600K at 5.0 GHz and almost same RAM freq. using your old track:

cache.thumb.png.8592884fbe0728021c9a25e20589ad5d.png

 

Only 1 fps difference. But they are not on the same version of the game. So apparently no influence.

 

So, for same freq, maybe there will not be difference with the i9-990K, or i7-9700K or i5-9600K, despite they have different caches:

 

2018-10-08-12_23_15-Intel-Announces-9th-Gen-Core-CPUs_-Core-i9-9900K-8-Core-i7-9700K-i5-9600K.png.791f598edbdeb76ca82cc04d332ac474.png

Share this post


Link to post
Share on other sites

Results from my rig using your standardized test procedure, IL2 version 3.006:

  • 8086k @ 5.1ghz, AVX -1, Ring Ratio 48.
  • DDR4-3200 16-18-18 memory clocked at 3466, same timings.
  • EVGA RTX 2080 XC Ultra at stock settings
  • CPU Mark 16516, single threaded 3084
  • Approximate peak GPU usage 83% but averaging more like 60% during most of the run.
  • Migoto mod disabled.
  • SteamVR in use (standard) not Open Composite.
  • IL2 FPS for test track: 72.233, 71.483, 71.225, 70.433

I tweak my setup differently for my own gaming but of course that would make the results not really comparable.

Share this post


Link to post
Share on other sites

Thank you Alonzo for bringing the first tests with a 2080 card and the 3.006 version. I have added them to the table.

Just for completeness of the data, which is your MoBo?

 

Since the v3.006 is has reduced the fps significantly (6fps in my rig), I have put the v3.005 results apart (at the bottom) and I will not mix those results for the correlation calculation.

 

As you can see your new GPU is infra-utilized with the Rift since the limiting factor is CPU.

But you can increase your SS since you have a a large margin in your GPU till 100%.

 

One question, apart from SS, have you seen any other graphics setting option which is loading only your GPU? (so you can exploit that)

 

Share this post


Link to post
Share on other sites
1 hour ago, chiliwili69 said:

Just for completeness of the data, which is your MoBo?

 

One question, apart from SS, have you seen any other graphics setting option which is loading only your GPU? (so you can exploit that)

 

Motherboard is an MSI Z370 SLI Plus. 10-phase power delivery and overclocks like a champ.

 

I think that SS and AA are the two main "GPU only" settings that you can pump upwards. The problem with SS is that although it makes the game look peachy, it can actually make enemies harder to spot because it smushes pixels. So I'm still experimenting.

 

Another option that increases both CPU and GPU is actually Lefuneste's 3D Migoto mod. He has put a lot of work into the mod and it fixes a bunch of broken shaders in game, for example you can see planes properly in front of clouds rather than have them disappear/flicker. But overall the mod seems to use quite a bit of GPU power for its effects, I was seeing ~20% difference in some runs. I'm certainly not criticizing the mod, it's great and gives zoom and fixes graphics, but it's actually quite a hit in performance.

Share this post


Link to post
Share on other sites

i have the lenovo explorer what resolution i have to set in steamvr to be in the baseline?

In the meantime i upload the passmark of my rig (updated cpu)

 

 

image.png.01a8ff16e346ccce94f181625d68a186.png

 

image.png.98a8a41ec1cc31b4dd317335cd789468.png

Share this post


Link to post
Share on other sites

anyway i have run the test with my steamvr settings per eye at trying to much as best as i can the other VR HMD i have seen in the examples (as said on the above post i have the lenovo explorer which has higher res then rift and vive but sligtly less then odissey or vive pro and has a square screen 1440x1440 per eye)

 

image.png.1e83df13edec7cf533ca77e541b08372.png

 

these are the results

 

2018-11-04 00:23:45 - Il-2
Frames: 6711 - Time: 120000ms - Avg: 55.925 - Min: 42 - Max: 91

2018-11-04 00:27:50 - Il-2
Frames: 6890 - Time: 120000ms - Avg: 57.417 - Min: 41 - Max: 91
 

 

my specs are the following

 

i9 9900k no OC (enabled MCE in bios only which took automagically all cores at 5GHz), MSI MAG Tomahawk z390 motherboard, gskill dd4 16gb at 3200Mhz, Asus Turbo GTX 1080ti (oc at 1987 maximum)

Don't know if i can be satisfied or not....but there is no better CPU that money can buy.....

Edited by Virus*

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×
×
  • Create New...