Jump to content
chiliwili69

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

Recommended Posts

Well that certainly looks like 4.5. It shows the multi running at 45. It would seem you have turbo kicking in.

 

I personally use AIDA 64, and have it set to display in my Logitech Keyboard LCD display.

Share this post


Link to post
Share on other sites

I also find it strange your il2 fps went down with the oc, especially after passmark verifying a performance increase and that you got it all right.

 

But if the current settings give you the higher performance in il2. Keep them!

 

As I understand it the Intel turbo will bump only one core at a time to 4.5ghz. While that oc you tried allowed them all to go there.

 

In your cpu-z screenshot, It's showing the speed of core #0, right click that value and you can see them all and at what speed they are running. You can probably have that open while running the benchmark. Though it could affect performance results.

 

If you want to dig deeper in to it all, this is a very good guide for everything you would want to know regarding your Asus bios and the 7700k overclocking.

 

http://edgeup.asus.com/2017/01/31/kaby-lake-overclocking-guide/

 

A lengthy prologue but worth it.

Edited by a_radek

Share this post


Link to post
Share on other sites

Thanks for the tip and info.  I checked by right-clicking in CPU-Z and it showed all 4 cores running at 4.5.  I'll take a read in that guide and see what I can figure out.

 

In the meantime, I ran another QMB with IL-2s vs. Ju-52s last night - forgot to turn on Fraps beforehand, but it was amazing looking and felt very smooth with no hint of motion sickness/discomfort.  Downed 3 of the 4 Ju-52s and damaged the 4th before running out of ammo - hung around to watch my AI buddies try to take out the 4th one, but saw him make it back to land at his airfield trailing fuel and smoking all the way back.  Then I managed to find a friendly airfield and land my IL-2 almost perfectly (ran about 15 feet off the end of the runway 'cause I hadn't remembered to assign brakes to anything when I re-arranged my controls for VR use).  After all these years of waiting for BoS to get to the point of my being able to enjoy it, these last couple of updates and the acquisition of my Rift have really done the trick.  I knew it would happen eventually, and with BoK and career mode almost here it'll just keep getting better and better - can't wait!

Share this post


Link to post
Share on other sites

 

 

After all these years of waiting for BoS to get to the point of my being able to enjoy it, these last couple of updates and the acquisition of my Rift have really done the trick.

 

Exactly. VR is really a game changer for BoS. It simply makes you believe you are inside the plane, even if resolution is not perfect it is enough for many of us.

I remember the first time someone shot me while flying, I could even feel the bullets in my body!. I was much more afraid of being shot that in a monitor.

After missions in VR, when you remove the Rift, you have the feeling of have been in a real war scenario, being part of it.

Share this post


Link to post
Share on other sites

 

 

Chiliwili, this thread was your initative and you did 99% of the work here. I can help out with the monitor test if needed (benchmark procedure would be simpler) And I do think it would belong better as a sister-thread in the general hardware forum, with a link to this one

 

No, I have just thought it would be great, and then Balapan did all the work and shared the test that we re-used to create the procedure. The work is done by the community, contributing with their test to help decide others with their upgrades. We all just want to spend the right amount of money for a pleasant experience, but no more. 

 

I have added the monitor procedure at the bottom of the first post. You can run the test now in VR and in monitor.

 

Thank you for your contributions. :drinks:

Share this post


Link to post
Share on other sites

Exactly. VR is really a game changer for BoS. It simply makes you believe you are inside the plane, even if resolution is not perfect it is enough for many of us.

I remember the first time someone shot me while flying, I could even feel the bullets in my body!. I was much more afraid of being shot that in a monitor.

After missions in VR, when you remove the Rift, you have the feeling of have been in a real war scenario, being part of it.

 

Amen to that!

 

I am currently flying Ivan's War Campaign, and having a blast. Flew a mission protecting some IL-2's, and had to engage some pesky 109 F4's that did not want to see the IL-2's make it to target.

 

I was behind one , he was banking left and I was laying cannon rounds on his left wing, that thing snapped off and went tumbling over my head and found myself trying to duck LOL.

 

When I made it back to my airfield I found myself opening the canopy to try and get some fresh air.

 

Man, flying in VR is so awesome.

Edited by dburne

Share this post


Link to post
Share on other sites

Can you guys help me with a little consensus regarding the other graphics settings that work best for you in VR?

Specifically cinematic camera on/off, SSAO on/off, HDR on/off, grass setting, distant view setting, AA on/off, overall graphics quality setting (low to ultra), SS setting, vsync on/off,

and anything else along these lines. The VR section of the forum has been extremely helpful for me for various settings, but i am having a hard time compiling the consensus of these settings in one place.

 

My system is Intel I-7 7700, not overclocked because I don't know what I am doing with that, 16 gb RAM, EVGA 1080, oculus rift, Windows 10 64 Bit.

 

What has worked for me has been cinematic camera on, SSAO on, HDR on, ultra grass setting, 4X distant view, AA off, graphics quality ultra, SS at 1.1, ASW off, vsync off.

I get a smooth appearance at these settings, but it is a little less smooth at SS 1.5 even if I turn graphics and other features to lowest settings.

Sometimes, even if I get a slighly higher framerate with lower graphics settings with SS at 1.3-1.5, the overall appearance seems smoother at SS 1.1 with the ultra settings in everything.else.

I just want to optimize what I have, which I feel is a good system, but I want to draw on you guys' experience to get the most out of it.

 

Thank you for your help!

Edited by BMW801

Share this post


Link to post
Share on other sites

What has worked for me has been cinematic camera on, SSAO on, HDR on, ultra grass setting, 4X distant view, AA off, graphics quality ultra, SS at 1.1, ASW off, vsync off.

 

 

If that works best for you, then you should probably stick with it.

 

Personally, I keep cinematic camera, SSAO and HDR off (with my monitor as well as with VR), because I don't like the way it looks with them on.  Also, I run AA=4 and SS at 1.7 (in SteamVR settings), because I hate jaggies and shimmering and that looks best to me and my new rig gives me good results at those settings.  Otherwise, I'm with you on ultra grass setting, 4X distant view, graphics quality ultra, ASW off, vsync off - it looks better to me that way and it doesn't seem to exact much of an FPS penalty compared to lower settings, at least on my rig.  Screenshot of in-game graphics settings is below:

 

post-12147-0-66685100-1502566969_thumb.jpg

Edited by TG-55Panthercules

Share this post


Link to post
Share on other sites

I run the following:

Cinematic Camera, SSAO and HDR off.

Ultra Graphics Settings

Terrain x4

Sharpen on the Filter

AA 2x

SS at 1.7 ( set in Oculus Tray Tool).

ASW Off

Share this post


Link to post
Share on other sites

My results for the Balapan Test for VR only:

 

Three runs to get an average.

 

 

2017-08-13 17:05:39 - Il-2
Frames: 4679 - Time: 60000ms - Avg: 77.983 - Min: 44 - Max: 90
2017-08-13 17:10:51 - Il-2
Frames: 4428 - Time: 60000ms - Avg: 73.800 - Min: 44 - Max: 90
2017-08-13 17:12:43 - Il-2
Frames: 4515 - Time: 60000ms - Avg: 75.250 - Min: 44 - Max: 90
 
System Specs in Sig.
 
edit- add passmark results for peoples comparisons.
 
Passmark results:
CPU Mark                           12340
CPU Single Threaded         2825
Edited by Wraithzlt

Share this post


Link to post
Share on other sites

Thank you Wraithzit, added to the sheet. Your avg fps is very much aligned to the one I had when I used 1866MHz memory. As you might have read form this thread, I also had a 4790K@4.7GHz with 1866MHz RAM, then upgraded the RAM to 2933MHz and I increases 8 fps my average. a_radek and Ganeshka were also reporting nice increments in fps when going to higher RAM speeds.

Share this post


Link to post
Share on other sites

 

 

Can you guys help me with a little consensus regarding the other graphics settings that work best for you in VR?

 

My usual graphics settings are very much aligned to the VR test procedure except for mirrors and landscape detail

 

SS at 1.7 (in Steam VR)

Cinematic camera: off

Graphics "HIGH".

Screen resolution: 1024x768

Full screen: off

Enable VR: ON

Multi GPU: off

Vsync: off

 

SSAO: off

HDR: off

Sharpen: ON

 

Mirrors: (depending on missions)

Distant landscape detail: x2

Grass quality: normal

Landscape filter: Sharp

Target FPS: off

Dynamic resolution: Full

Antialiasing: x2

Gamma correction: 0.9

Share this post


Link to post
Share on other sites

 

 

My system is Intel I-7 7700

 

Is 7700 or 7700K?

 

OC is only possible with 7700K.

 

If it is 7700K you will benefit a lot going to higher clock speeds. I also thought OC was going to be very complex, but with the MSI Command Center was very easy. You only need a good CPU cooler and spend a couple of hours reading the guides. There is a thread in this forum about that.


 

 

SS at 1.7 ( set in Oculus Tray Tool).

 

So you mean SS=1.7  (in OTT), which is equivalent to 2.9 SS in SteamVR?

 

Or you mean SS=1.3 (in OTT), which is equivalent to 1.7 SS in SteamVR?

Share this post


Link to post
Share on other sites

 

 

So you mean SS=1.7  (in OTT), which is equivalent to 2.9 SS in SteamVR?

 

Or you mean SS=1.3 (in OTT), which is equivalent to 1.7 SS in SteamVR?

 

1.7 in Oculus Tray Tool.

Share this post


Link to post
Share on other sites

Thank you for your responses and help!

It is an i7 -7700k CPU at 4.2 GHz.

I'll check out the MSI command center.

Share this post


Link to post
Share on other sites

 

 

It is an i7 -7700k CPU at 4.2 GHz. I'll check out the MSI command center.

 

I think MSI Command Center would only work on MSI Motherboards, but not sure.

If you have another MoBo manufacturer like ASUS or Gigabyte, they also provide similar tools to do OC.

 

In any case, I will be curious to know what benchmark you achieve at 4.2GHz and once it is OC'ed. 

Share this post


Link to post
Share on other sites

 

 

1.7 in Oculus Tray Tool.

 

Yeah!, you have to amortize your GPU!  :P

 

I saw that initially you used 1.3 or 1.5, but with the new GPU went to higher SS.

 

I think that, once the CPU/RAM delivers a good benchmark (let´s say above 80 in this VR IL-2 benchmark) at SS=1.0, then people can increase their SS until the fps dramatically drops.

 

For my 1070 card that limit is around 1.4 (in OTT, or 2.0 in SteamVR), but I will do some more testing when to determine that.

 

Have you tried to run this benchmark with your upgraded rig?

Share this post


Link to post
Share on other sites

 

Have you tried to run this benchmark with your upgraded rig?

 

No too busy campaign flying!

 

:salute: 

Share this post


Link to post
Share on other sites
No too busy campaign flying!

 

Came on Don! it is no more than 15 min

you shouldn´t be shy, we are not comparing our co..., just our rigs   :blush:  

Edited by chiliwili69

Share this post


Link to post
Share on other sites

:blink: ...gaaaaaaaaaaaaaaaaaaaaaaaayyyy  :hunter:

 

 

I will pull my .... data out once my i7 7700k arrives. With and without overclock will be compared, as well as different RAM speeds.

Edited by 1CAG_Fenris_Wolf

Share this post


Link to post
Share on other sites

While we already have benchmarks on higher end cpu's. The more interesting ones are from the older rigs.

 

My guess is many if not most are running things like Sandy bridge (i5 2500k) And are shying away from VR as they don't feel like replacing a perfectly good gaming computer only to try it.

 

My second guess is that those older CPUs, especially overclocked and with some faster ddr3 would do perfectly fine. But we can't really know until we have the numbers.

Edited by a_radek

Share this post


Link to post
Share on other sites

 

 

 ...gaaaaaaaaaaaaaaaaaaaaaaaayyyy      I will pull my .... data out once my i7 7700k arrives. With and without overclock will be compared, as well as different RAM speeds.

 

Lol. No, no gay (btw nothing against them), just wanted to press Don a bit.

 

As Radek indicated it is also important to have benchmarks for mid-range PCs with low-med-top GPU, and also top-range PCs with low-med-top GPU.

 

Please, this is foor all people reading this, don´t wait for your PC upgrade to report your benchmark. All data is valuable.

Share this post


Link to post
Share on other sites

2017-08-16 11:39:31 - Il-2
Frames: 3145 - Time: 60000ms - Avg: 52.417 - Min: 43 - Max: 88

 

SS of 1.3 set in OTT.

Share this post


Link to post
Share on other sites

I have no VR headset so all measurements are made for full HD monitor.

 

Settings from first message in order they change some day:

 

 

Under Graphics settings choose "HIGH".

Screen resolution: 1920x1080
Full screen: off
Enable VR: off
Multi GPU: off
Vsync: off

SSAO: off
HDR: off
Sharpen: ON

Mirrors: off
Distant landscape detail: normal
Grass quality: normal
Landscape filter: Sharp
Target FPS: off
Dynamic resolution: Full
Antialiasing: x4
Gamma correction: 0.9

 

 

 

Version 2.011

 

CPU i7 3770K 

1-core active: 4.9GHz

2-core active: 4.9GHz

3-core active: 4.7GHz

4-core active: 4.7GHz

Voltage 1.350 mV

 

DDR3 1600 16GB

 

Passmark

10332 full

2696 single-thread

 

Measurements averaged of 3 runs for each GPU

 

1060 3GB: 8749, 84, 196, 146 (frames, min, max, avg)

1070 8GB: 8921, 112, 207, 148

1080 8GB: 8882, 115, 204, 147

1080 Ti 11GB: 8969, 116, 212, 149

 

post-2537-0-96829200-1502899989_thumb.png

 

Now my IMHOs

 

Balapan's track is good to compare rigs, but I suppose track is lacking of physical calculations. In online battles at Berloga server FPS almost never locks at 120 (I have 144 Hz BenQ running at 120 Hz) with frequent drop downs to 60-70 and this is with medium settings, unlike in benchmark.

 

There is no current hardware which allow locking FPS at 120 and playing with low motion blur displays

 

Obviously, CPU single thread performance is bottleneck. CPU manufacturers cannot improve this much due to physical limitations. Multi-core is our present and future.

 

IL-2 is indeed multithreaded but there is one thread which works most of the time and limits overall performance

Edited by TUS_Samuel

Share this post


Link to post
Share on other sites

Finally got round to Ryzen benchmark. I think something's very odd with my system as my performance should be much higher and my CPU utilisation on the most-used core sits around 40-60% while GPU is at 30-50%. I'm downloading the IL2 game standalone so I can test without Steam version.

 

Chilli, my old I5 4670K benchmarks were totally wrong, I wasn't using FRAPS correctly so you can ignore them.

 

Have Ryzen 1700 @ 3.65GHz, 16GB 2933MHz RAM, GTX 1070 Asus B350 Plus MB

 

Screen results

2017-08-17 10:33:40 - Il-2
Frames: 5460 - Time: 60000ms - Avg: 91.000 - Min: 73 - Max: 119
 
VR
 
2017-08-17 10:57:58 - Il-2
Frames: 2674 - Time: 60000ms - Avg: 44.567 - Min: 41 - Max: 51
 
Low graphics settings
2017-08-17 11:00:24 - Il-2
Frames: 2840 - Time: 60000ms - Avg: 47.333 - Min: 43 - Max: 80
 
What's odd is that in-game I've seen a massive improvement, often even getting 90fps taxiing and most of the time in the air and drops to 45 are normally only flying close to ground or with lots of AI/Players around. (Low graphics)
 
So I'm stumped really. The experience is much better than with my I5 but benchmark says otherwise. My GPU is not remotely being utilised (40% avg) and CPU is hardly being taxed (50% avg) so why isn't the game pushing my hardware? It's like I've got throttling turned on somehow.
 
Passmark benchmark gives me total score of 4674.9 and CPU of 13883.3 single thread 1727
Edited by AceVenturi

Share this post


Link to post
Share on other sites

I've had the same with my 1600x at 4.0 GHz and a GTX 1080TI. BoS just will not run at 90 FPS despite everything being below 50% usage in the Balapan track. In actual play, it will hit 90 fps at times when up off the ground a little but even then, it's mostly 45. Passmark single thread rating is ~2130. Other games aren't a problem and the really weird bit is DCS actually runs  better for me in VR than BoS does. That's one hell of a reversal from 2d results.

Share this post


Link to post
Share on other sites

 

 

Frames: 3145 - Time: 60000ms - Avg: 52.417 - Min: 43 - Max: 88

 

Thanks Don for reporting. I would expect a higher benchmark from your rig.

 

Do you use 4K skins?  radek used that and had a big fps impact.

 

Try to run the Passmark becnhmark in your CPU a see what you have (including single thread) 

Share this post


Link to post
Share on other sites

 

 

Balapan's track is good to compare rigs, but I suppose track is lacking of physical calculations.

 

Many thanks Samuel for running this with several GPUs! (BTW, how do you have so many GPUs and no VR?)

 

Congratulations also for your good single thread performance with the 3770K, it is well above average tests.

 

This test doesn´t take into account the online playing, since this depend very much on the online servers, internet bandwidth, etc. We just wanted to analyse this from the user PC side.

 

As you guess, there is one thread of IL-2 which is the one responsible for the overall performance. Splitting this thread in multiple threads would not be an easy task.

 

Fomr you test I clearly see that GPU budget doesn´t translate to more fps. It is wiser to spend that budget in CPU/RAM.

 

Also, if you can try your rig in VR, you would obtain about 70-75 avg fps in the benchmark (just a ballpark rule of dividing by 2 the monitor performance)

Share this post


Link to post
Share on other sites

Thanks Don for reporting. I would expect a higher benchmark from your rig.

 

Do you use 4K skins?  radek used that and had a big fps impact.

 

Try to run the Passmark becnhmark in your CPU a see what you have (including single thread) 

 

No 4k skins, all default stock game.

 

I was a little surprised as well, especially as my performance is so good in flying the campaigns.

Edited by dburne

Share this post


Link to post
Share on other sites

 

 

1060 3GB: 8749, 84, 196, 146 (frames, min, max, avg) 1070 8GB: 8921, 112, 207, 148 1080 8GB: 8882, 115, 204, 147 1080 Ti 11GB: 8969, 116, 212, 149

 

I have taken a closer look if the graph you showed. 

 

Normally the fps is in the 140-160 fps range, and the 3 cards show almost the same performance. So CPU is bottleneck.

 

Only when the balapan track show no planes or smoke, the fps rise to 190-210 range and here GPU is bottleneck, so every GPU separate from the other.

 

And then, there is a general big drop when the last bomber is damaged and fire/smoke appear massively on scene. Here only the 1060 lose the trail of the other three cards.

 

I think the Balapan track was created with all sort of things (clouds, stalingrad map, multipleplanes, maybe turbulent wind, etc). I will ask him how it was created.

Share this post


Link to post
Share on other sites

Also found Dons result peculiar. So tested running 1.3 in tray tool instead of 1.7 in steam and got a significant drop in fps. Did not doublecheck resolution in occulus monitor, and was testing other things at the same time so can't be sure all was correct. Will investigate further another evening, short on time for a few days now.

Share this post


Link to post
Share on other sites

Another thing I have found since installing Fraps I now have severe mouse lag when I first launch BoS. Upon getting to the log in screen, it will take around 2-3 minutes before my mouse becomes responsive - whether Fraps is running or not.

 

Going to uninstall it and see if that corrects the issue.

Share this post


Link to post
Share on other sites

 

 

Going to uninstall it and see if that corrects the issue.

 

Ok uninstalled Fraps, mouse behavior back to normal now whew.

Share this post


Link to post
Share on other sites

IL-2 is indeed multithreaded but there is one thread which works most of the time and limits overall performance

 

Here is an illustration

 

post-2537-0-48025900-1502981707_thumb.png

 

You can see all IL-2 threads on this graph while playing back Balapan's track. Green at the beginning is track loading, does not matter. Then it goes actual playback, red graph shows about 90% CPU usage by the bottleneck thread with maximum peak at 98% (around the last He-111). 

 

Also you can count total CPU usage from this by summing average CPU usage and dividing by core count. In my case it is (90% red + 20% purple + 10% green + 10% other) / 4 cores = 32% which corresponds with task manager.

 

That's why you see only 30-40% total CPU usage, while FPS is low

Share this post


Link to post
Share on other sites

I added a link (Passmark test) to the procedure, so people can also report their individual passmark scores.

 

BTW, there are a page where the Single-thread CPU are sorted:

 

https://www.cpubenchmark.net/singleThread.html

 

and also another page with more info that your can order as you wish:

 

https://www.cpubenchmark.net/CPU_mega_page.html

Share this post


Link to post
Share on other sites

 

 

That's why you see only 30-40% total CPU usage, while FPS is low

 

SAMUEL!!  Thank you for unveiling something that everybody was suspecting but didn´t proof with data.

 

It was always the typical complaint (CPU is not loaded!) that we had (including myself).

 

Just curious, what tool did you used to obtain and record that data?

Share this post


Link to post
Share on other sites

The discussion in another thread about GPU upgrades for VR (here: https://forum.il2sturmovik.com/topic/30567-gpu-upgrade-advice-vr-please/  ) got me curious so I went in and ran another test on my 7700K - ST passmark result of 2647, running at 4.2 (I think - can't tell if it's running in turbo mode (4.5) or not - CPU-Z seems to think it is, but Passmark seems to think it's not) - I decided to make one more try at OC'ing to 4.5 and see what happens.

 

Tried using the ASUS utility that came with my MB to OC my CPU - it runs a series of iterative overclocking attempts in .1 increments, and got it up to 5.0 before I stopped it.  Ran a new Passmark test at 5.0, and the ST index/result jumped to 2937.  Then I decided to run another test with the VR benchmark (from 1st post above) to see if there's any practical benefit at this level.

 

Surprisingly (to me), the test on the VR benchmark (from first post above) with my CPU OC'd to 5.0 did not result in any improvement at all compared to running it at stock (either 4.2 or maybe 4.5 turbo - can't tell whether turbo was used or not) - actually a slight reduction at 5.0, but so close as to be within the realm of standard deviation.  But if it's correct, it would seem to indicate a point of significantly diminishing returns to overclocking CPU at this point.  

 

Really curious as to why there wasn't at least some small but noticeable improvement from such a large overclocking.  Didn't notice any unusual fan activity or temperature spikes so I don't think there would have been any throttling kicking in, although I don't know enough about how all this OC stuff works to be sure about anything at this point.

Edited by TG-55Panthercules

Share this post


Link to post
Share on other sites

SAMUEL!!  Thank you for unveiling something that everybody was suspecting but didn´t proof with data.

 

It was always the typical complaint (CPU is not loaded!) that we had (including myself).

 

Just curious, what tool did you used to obtain and record that data?

Looks like Performance monitor in Windows. 

 

I did what I thought is the same test using %processor time and got these two results, one with screen test and one in VR.

 

In VR my most used thread only gets to 60% average whereas on screen it goes up to 75% average. Weird but in either case it's not maxing out the CPU.

 

Samuel what counter did you use in Perfmon to get your graph?

post-77541-0-75034800-1503014022_thumb.jpg

post-77541-0-75518000-1503014043_thumb.jpg

Edited by AceVenturi

Share this post


Link to post
Share on other sites

0. This is perfmon. It is a windows utility. You can run it as simple as press Win+R and type perfmon, press Enter.

 

AceVenturi, try this

1. Run IL-2 and minimize

2. "Add Counters" -> "Thread" -> "% Processor Time" -> type "IL-2" in textbox -> select all thread instances -> click "Add"

 

Difference is that on your 8 graphs there are usages of each core and on my graphs there are CPU usages for each IL-2 thread. However core usage is also a very interesting information.

 

Below is just my interpretation and only devs can say exactly what is happening. However

CPU load in VR may drop due to a fixed time at which a frame should be ready. If a frame could not be delivered on time then it gets ready whithin next time frame and CPU sleeps for the rest of time. This cannot happen on screen because FPS is not limited and as soon as CPU delivers previous frame it starts processing the next, this gives full CPU utilization.

You can try changing pre-rendered frames in NVIDIA control-panel and see what happens. I think core utilization should be higher and FPS may rise in VR.

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...