Jump to content
chiliwili69

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

Recommended Posts

 

 

1440p monitor result: AVG: 108 Min: 73 Max: 148

 

I was going to reply beforebut has no time.

The performance you were having with the RX480 1440p was good, but in order to be comparable with previous test you can run it in a monitor at just FullHD resolution (1080p). If you have above 140 Avg, then you should be OK for VR (just a ballpark number).

 

But there should be something strange for the VR in your case.

 

The Ryzen 5 1600X is a powerful CPU, it has a Passmark of 13137, higher than 7700K (12163). But a Single Thread Rating of 1944 (https://www.cpubenchmark.net/cpu.php?cpu=AMD+Ryzen+5+1600X&id=3000 ), but 7700K has a Single Thread Rating of 2585.

 

Now your 1080Ti is the top one and your memory is at 2666MHz.

 

You can try to OC to 4.0Ghz to see if you get an improvement.

 

What MoBo are you using?

 

Have you tried to run the Oculus compatibility check?

Share this post


Link to post
Share on other sites

It isn't a CPU bottleneck, no single thread is anywhere near maxed in that track and it isn't hopping threads either. I've looked into it pretty deeply, and something is forcing an ASW like thing on. It is not actually ASW as ASW has a very visable prop flickering/jittering effect on objects on the far side of it (ie planes). Hitting Ctrl + 1 removes that effect, so I know ASW for Oculus is off (turning ASW back on with ctrl + 4 brings the flickering back). Despite that, frame times are still stepping between 11ms and 22ms, corresponding to 90 fps and 45 fps respectively. I even looked off to the side in that track to try to balance it out at 60 fps or so and had some success, but all that was going on is it was bouncing between the two steps, it didn't give a whole bunch of 16ms results as you'd expect, just a bunch of alternating 11 and 22 ms results. I checked steamVR settings to see if the problem is in there but all 3 options for ASW like effects I could find were turned off ("enable always on reprojection" in "Developer", "Allow asynchronous reprojection" and "Allow interleaved reprojection" both in "Performance"). I even tried reinstalling both Oculus and SteamVR without effect. I'm at a complete loss as to what to do about it. I don't have v-sync or anything like that turned on anywhere, only other thing I could think of.

Share this post


Link to post
Share on other sites

OK - decided to give this benchmarking thing a try.  I'm pretty sure I followed all the directions, and wound up with the following results:

 

Initial test suite with GTX 980 (PC specs below)*:

 

Test with VR and specified VR test settings:  Frames: 3461 - Time: 60000ms - Avg: 57.683 - Min: 35 - Max: 63

 

Test with monitor and specified VR test settings:  Frames: 7847 - Time: 60000ms - Avg: 130.783 - Min: 100 - Max: 163

 

Test with VR (1.7 SS) and my usual monitor settings**:  Frames: 3331 - Time: 60000ms - Avg: 55.517 - Min: 30 - Max: 62

 

Test with monitor and my usual monitor settings**:  Frames: 7014 - Time: 60000ms - Avg: 116.900 - Min: 71 - Max: 153

 

*PC specs:  Intel Core i7-4770K @3.5 GHz; ASUS ROG Maximus VI MB; 16GB RAM (DDR3 1600MHz)

EVGA GTX 980, 4 GB; Realtek ROG SupremeFX audio
Windows 10 Pro 64-bit

 

**usual monitor settings (differences from specified VR test settings) = 1920x1080; Ultra; SSAO off; mirrors = complex; distant landscape detail = 4; grass quality = distant; Antialiasing = 4.

 

 

[EDIT]

 

Second test suite, after upgrade to GTX 1080TI (same specs as above except for Video card):

 

Test with VR and specified VR test settings:  Frames: 4554 - Time: 60000ms - Avg: 75.900 - Min: 61 - Max: 94

(gain of only 18 FPS on average)

 

Test with VR (1.7 SS) and my usual monitor settings: Frames: 4017 - Time: 60000ms - Avg: 66.950 - Min: 54 - Max: 81

(gain of only 11 FPS on average)

 

Test with monitor and my usual monitor settings:  Frames: 7490 - Time: 60000ms - Avg: 124.833 - Min: 97 - Max: 160

(gain of only 7 FPS on average)

 

Needless to say, I'm incredibly underwhelmed with these results.  Guess I'll try to find some benchmarks to run to see if the new card is underperforming generally or if maybe my CPU/RAM is more of a bottleneck than I thought :(

Edited by TG-55Panthercules

Share this post


Link to post
Share on other sites

I understand the problem is not in BOS but in the Rift with SteamVR games. Have you tried other SteamVR games apart from DCS?

There is a tool called FCAT VR to analyze performance, I still never used it but could bring some light here.

 

There are several games benchmarked with FCAT VR: http://www.roadtovr....rx-480-fcat-vr/

 

You can post this problem in the Oculus forum, there will be many Rift users with AMD platforms.

Share this post


Link to post
Share on other sites

Right, screw it, it's never going to do anything but 45 or 90 fps.

System:

CPU: R5 1600x (4.0 GHz)

GPU: EVGA GTX 1080 TI SC2 ICX

RAM: 16GB DDR4 2667 MHz

Mobo: MSI B350 Tomahawk

OS: Win 10 Home

 

post-13947-0-12789000-1501867311_thumb.gif

Share this post


Link to post
Share on other sites

*usual monitor settings (differences from specified VR test settings) = 1920x1080; Ultra; SSAO off; mirrors = complex; distant landscape detail = 4; grass quality = distant; Antialiasing = 4.

 

Rig specs are in signature.

 

Not sure why the VR results are so low, or why there's so little difference between the VR results with all my usual settings cranked up.  Have been considering upgrading my GTX 980 to a 1080 or 1080Ti, but I'm not sure how much good that would do.  Maybe upgrading my CPU/RAM/MB would make more sense?

 

Lower AA to 2. You might can up SS somewhat then as well, not sure where you are running it or if you are.

 

That CPU of yours running at stock? It would be a great chip to overclock some and easily done so , if you have a decent cooler. I am running a 4820K at 4.5 Ghz nicely.

 

A 1080 Ti should give a significant boost in VR over a 980 card. Whether the amount of boost is worth the expense to you might be another thing. I went from a 1080 to a 1080 Ti and do not regret it.

Share this post


Link to post
Share on other sites

Lower AA to 2. You might can up SS somewhat then as well, not sure where you are running it or if you are.

 

AA was set at 2 (per the first post's specs) for the first set of tests - I put it back to 4 for my second set of tests - and essentially no difference in results.  Running SS at 1.7 in Steam VR (per the first post's specs) for both sets of tests.

 

 

That CPU of yours running at stock? It would be a great chip to overclock some and easily done so , if you have a decent cooler. I am running a 4820K at 4.5 Ghz nicely.

 

Yeah - I messed around with OC'ing the CPU when I first built my rig a while back (have a nice liquid cooler that should be able to handle it), but wasn't seeing much in the way of results and wasn't sure I wasn't seeing some stability problems, so I just backed off.  I guess I could try it again, but I'm not sure it would squeeze out enough FPS increase to make a very significant difference.

 

What the heck - might as well see what happens :)

 

[EDIT] - well, now I remember another reason I stopped my OC experiments - there are way too many unfathomable choices in the BIOS settings of this MB.  I thought I remembered something a lot simpler - maybe some sort of OC software I used to have but have gotten rid of since?  Anyway, I did switch it to turbo mode, and also OC'd my GPU a bit.  Gained about 8-10 fps on my monitor, but only gained about 2 fps on my VR tests.    Been thinking maybe it was time to build another PC - I wonder what sort of improvement I'd likely see from both a PC and a GPU upgrade?

 

[EDIT] Ran a test with dropping SS from 1.7 to 1.3 - gained about 11 FPS on average:

 

SS - 1.7:  Frames: 3331 - Time: 60000ms - Avg: 55.517 - Min: 30 - Max: 62

 

SS = 1.3:  Frames: 3993 - Time: 60000ms - Avg: 66.550 - Min: 41 - Max: 75

Edited by TG-55Panthercules

Share this post


Link to post
Share on other sites

Well, decided that since I would be buying a new video card for my new rig (if I decide to build one) I might as well go ahead and get it now and try it with my current rig, so I picked up a new GTX 1080TI tonight.

 

As noted in my edited post (#83) above, the results were massively underwhelming.  Anybody got suggestions for a good benchmark program to run to see if the card itself is underperforming?

Edited by TG-55Panthercules

Share this post


Link to post
Share on other sites

https://www.passmark.com/ offers a full PC benchmarking tool for free, you can check graphics only if you'd like and compare to the scores they list. Thinking of which, I should do the same.

 

Thanks - I gave that a try, ran the 3D graphics tests, and came up with a 3D Mark score of 15452, which seemed to be a bit higher than the average of such scores I was seeing on the web for the 1080 TI, so I suppose that means the card is doing OK.  I guess that probably means my CPU and RAM are somehow holding things back, or something else is going on that I don't understand - gotta do some more digging to figure out how people are reporting 90+ (guess one thing to try would be to turn everything down and see what happens, though I'm pretty sure I'm gonna hate the way that would look).

 

[EDIT]

 

OK - at 1024x768 with everything turned off/down to minimum in game, and SS=1.0, the results were as follows:

 

Frames: 6213 - Time: 60000ms - Avg: 103.550 - Min: 81 - Max: 130

 

Unfortunately, as I feared, it looked like crap - but in some ways not nearly as bad as I expected.  I guess tomorrow I might try starting to crank some settings back up and see if there's a sweet spot in here somewhere.

Edited by TG-55Panthercules

Share this post


Link to post
Share on other sites

My benchmark, according to instructions and settings in the first post in this thread.

Frames: 3674 - Time: 60000ms - Avg: 61.244 - Min: 43 - Max: 91

This result is the average of running the test three times.

If I turn off SSAO, I get:

 

Frames: 4214 - Time: 60000ms - Avg: 70.244 - Min: 43 - Max: 91

No SSAO gives me an extra 9 fps on average. While my cpu is very good, it seems my gtx1070 is struggling. Also my results don't match Chiliwilis, as he is running the same GPU. We are trying to figure out what's going on, but for now these results stand.

Specifiations:

 

Version 2.011

Steam VR SS:1.7

Motherboard: Asus Prime Z270-A

I7 7700k @4.8Ghz

Current passmark @4.8Ghz: 13487 
Current passmark Single thread @4.8Ghz: 2806
 

16 Gb ddr4 @3000

Gainward GTX 1070 gs

 

Occulus Rift

Edited by a_radek

Share this post


Link to post
Share on other sites

 

 

Test with VR and specified VR test settings:  Frames: 4554 - Time: 60000ms - Avg: 75.900 - Min: 61 - Max: 94 (gain of only 18 FPS on average)

 

I see you wrote Max:94

In VR, Fraps should not report more than 91.

Some weeks ago, I remember that Fraps were reporting a Max fps higher than 91 in VR. But the reason for that is that the track is only displayed at the monitor but not at the Rift. Since the Rift is not wear on during the test I didn´t notice that the scene was not represented in the Rift.

So if your test report more than 91 in VR, something wrong has been done with the rift.

I didn´t know what was the problem for that, I just restarted the PC an the problem disappeared. 

Share this post


Link to post
Share on other sites

 

 

Needless to say, I'm incredibly underwhelmed with these results

Your expectations are too high IMHO. You should look at the min FPS gains, not the average.

 

 

 

Test with VR and specified VR test settings: Frames: 3461 - Time: 60000ms - Avg: 57.683 - Min: 35 - Max: 63

 

 

Test with VR and specified VR test settings: Frames: 4554 - Time: 60000ms - Avg: 75.900 - Min: 61 - Max: 94

 

You almost doubled the min FPS. In VR 35 FPS is unplayable, 65 FPS isn't ideal but OK, depending on individual preference and tolerance.

  • Upvote 2

Share this post


Link to post
Share on other sites

Feeling a little pukey...

Hi all, stumbled on to this thread while searching for any useful information for trying to improve VR with a GTX970.  I've been having nausea issues in BOS and am trying to pinpoint the exact problem.

 

I don't get sea-sick, air-sick, no nausea on theme park rides, have flown in a multitude of aircraft as a passenger and once I even got to fly a Helicopter between tress up a riverbed and have never had any nausea issues.

 

I tried to run this test to try get a benchmark for what's going on and where, but I can't load the graphics test past the flight records screen.  I uncompressed the file and then deleted the compressed version.  Will it not load because of the graphics card I have? or could there be another issue somewhere?  I would like to do a bunch of tests to get a baseline and then try overclocking everything to see if it helps.

 

I bought the Summer Bundle as I have been waiting for VR implementation in a flight sim for a few years after not really playing much, but keeping an eye on developments.  I took advantage of the Summer Sale and picked up teh rift and Touch controllers.  While it is awesome in a lot of the Oculus store demos I have played, the GTX 970 is not up to scratch in BOS from what I am experiencing?

 

I may have to look at piling on some more debt to shell out for a 1080 to rapidly improve things as I won't be playing BOS for much longer...

 

Judging by what I can see in this thread:

 

-Minimum/Average FPS is the most important thing to look at since RIFT is capped around 90 on the Rift and the low frame rates are what bring on the nausea?

 

-There may be also be a strong correlation between the CPU and how many fps there are in BOS in VR?  If so, could I bring my frame rate up to perhaps acceptable levels with some overclocking of my GPU and CPU? Is there a correlation between RAM and Frame rate in BOS?

 

I'm constantly dealing with Nausea while playing in just a 2-3 quick mission with my GTX970 from anywhere between 80fps and often around 35-45 fps according to the frame counter, possibly even lower when I peek at when pushing backspace.  My wife tried out a bunch of the demo games, was loving them and then strapped into BOS and was instantly feeling sick, so I knew somethings not right.

 

The low quality settings almost make things worse with trees popping in and out everywhere and random parts of the tree being re-drawn on a fly by.  I have 20-10 vision which means I notice every little detail (means better than 20-20, ie. I can read pretty much the whole eye chart at the optometrists) although I am a bit farsighted so the headset may be straining my eyes a little at the same time.  I think both of these things may even be increasing the nausea...

 

Basically If I play on higher quality settings, I get some nausea even with low frame rates, lower quality settings I get even more nausea as the lack of clarity really adds to the nausea experience..

 

At this stage it looks like I can never be in a multiplayer online battle in VR which has always been the point of owning this game until I improve this situation.  Am I best to shell out for a 1080 and hope it's still good for next gen vr? Or perhaps save up for a next gen Volta card when that maybe gets released early next year?  Since I can't actually see what VR actually looks like in this game with a 1080 I feel like I can't understand of things will improve a whole lot.

 

Any advice would be appreciated..

Thanks in advance!

 

 

 

Share this post


Link to post
Share on other sites

 

Does Box crash when you attempt to load the track/recording or is it that you can't see the track in your list?

 

The 970 will do a 45fps but it will struggle trying to reach 90. I had one and tried the rift/Box using it. You will only reach 90fps at very high altitude or looking up at the sky. It is after all the minimum recommended Gpu according to occulus rift hardware requirements.

 

Cause of the nausea issues is as far as I understand it very subjective. Framerates could very well be it if your dropping down below 45fps regularly.

 

Chiliwili is using the same CPU you are and he documented his whole journey in this thread. From so so, to stellar performance.

 

We don't know how amount of ram affects frame rates. But we do know Box takes advantage of ram speed more than games do in general.

Share this post


Link to post
Share on other sites

 

 

Minimum/Average FPS is the most important thing to look at since RIFT is capped around 90 on the Rift and the low frame rates are what bring on the nausea?   -There may be also be a strong correlation between the CPU and how many fps there are in BOS in VR?  If so, could I bring my frame rate up to perhaps acceptable levels with some overclocking of my GPU and CPU? Is there a correlation between RAM and Frame rate in BOS?

 

In VR you should not going below 45 fps at any time, otherwise it will cause you nausea. That´s why the Rift refresh rates of the DK2 went to 75Hz and CV1 to 90Hz. Your brain is specially good at detecting delays higher than 20-25ms (or lower depending on people) between your head move and what you see in the world.

So, fps would be the biggest contributor to your nausea.

 

VR should be a pleasant experience (even more inside a plane). Nausea should not be there, as soon as you feel bad stop it.

Making tests with the baseline will help to you to determine what settings or hardware is limiting your BoS VR experience. That´s the purpose of the baseline. Again, the Balapan track we use for the test is very demanding (clouds, several planes in scene, a lot of smoke). My usual flights are not so demanding.

 

BoS is very dependent on CPU, specially in VR.

I had the same problem than you (low fps but never below 45), so upgraded to a 4790K (the top CPU for my socket) and then overclok and changed RAM speed. This is documented here:

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

More or less you would gain 4 to 1.5fps per 0.1 GHz OC increase. You have a liquid cooler, so for sure you can go higher than me.

 

What RAM speed do you have? (It is not at signature)

For RAM, I had about 0.8 fps increase per 100MHz. I gained 8 fps with my upgrade from 1866 to 2933MHz.

 

I would not go to a 1080 until you try to squeeze your CPU and GPU as much as possible.

In my case, a 4790K and a 1070 is more than enough for VR with SS at 1.7.

 

Regarding the problem with the test, Do you see the balapan "test1" record in the BoS recorded screen?

The zip contain a file and a folder. Both need to be at the "track" folder.

Have you tried to run the test in a monitor?

Or have you tired to record your own tracks and replay them?

Share this post


Link to post
Share on other sites

In VR you should not going below 45 fps at any time, otherwise it will cause you nausea. That´s why the Rift refresh rates of the DK2 went to 75Hz and CV1 to 90Hz. Your brain is specially good at detecting delays higher than 20-25ms (or lower depending on people) between your head move and what you see in the world.

So, fps would be the biggest contributor to your nausea.

 

VR should be a pleasant experience (even more inside a plane). Nausea should not be there, as soon as you feel bad stop it.

Making tests with the baseline will help to you to determine what settings or hardware is limiting your BoS VR experience. That´s the purpose of the baseline. Again, the Balapan track we use for the test is very demanding (clouds, several planes in scene, a lot of smoke). My usual flights are not so demanding.

 

BoS is very dependent on CPU, specially in VR.

I had the same problem than you (low fps but never below 45), so upgraded to a 4790K (the top CPU for my socket) and then overclok and changed RAM speed. This is documented here:

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

More or less you would gain 4 to 1.5fps per 0.1 GHz OC increase. You have a liquid cooler, so for sure you can go higher than me.

 

What RAM speed do you have? (It is not at signature)

For RAM, I had about 0.8 fps increase per 100MHz. I gained 8 fps with my upgrade from 1866 to 2933MHz.

 

I would not go to a 1080 until you try to squeeze your CPU and GPU as much as possible.

In my case, a 4790K and a 1070 is more than enough for VR with SS at 1.7.

 

Regarding the problem with the test, Do you see the balapan "test1" record in the BoS recorded screen?

The zip contain a file and a folder. Both need to be at the "track" folder.

Have you tried to run the test in a monitor?

Or have you tired to record your own tracks and replay them?

 

Sorry, RAM speed has been corrected and is in signature.

 

I will try the following when I get time to:

-Will check the files on the balapan test, if you have a screeshot of what your folder looks like when correct, that would help out big time! I will try to run test just on monitor, if it loads, will then try VR.

 

-Overclock 4790K, I haven't done this before, but there seems to be a lot of info to get this done online, if you can see any guides that worked well for you, I would appreciate a link if you have one.

 

-Overclock GPU

 

-Enable XMP profile in BIOS

 

-Will run test with Monitor only again and then try VR test.

 

Thanks for your help!

Share this post


Link to post
Share on other sites
if you have a screeshot of what your folder looks like when correct,

 

I am two weeks out of home, so I'm not able to run BoS and produce new screenshots.

 

Radek, could you send me that screenshot and I can update that in the procedure, so it is clearer? 

 

Regarding your RAM speed (1333), this is the lowest value your MoBo supports. You can go up to 3300 if you put new high speed RAM. (but it is difficult to find now DDR3 high speed RAM, I found mine at amazon.com, it is also not cheap either)

 

For the OC, every MoBo manufacturer has their own tools which facilitate the OC for newbies like me. Instead of BIOS, I used MSI Command Center. I suppose you'll have similar tools from Gigabyte. In the OC thread I mention you will find advices I followed from community. I never thought it was so easy.

 

I didn´t OC my GPU (I am happy with my fps so far) but it will be interesting to see how it helps in your 970.

 

Your MoBo has several PCIE slots for the GPU, verify that your 970 is in the optimal slot (the fastest)

Edited by chiliwili69

Share this post


Link to post
Share on other sites

Of course.

 

tracksfolder.jpg

 

I hope this helps. The "test1" folder, and the "test1.trk" file. Should be inside the il2 tracks folder.

Share this post


Link to post
Share on other sites

OK - finished my new PC build and ran some more tests.  Just to keep everything together, I've pasted in my previous sets of test data results below:

 

 

Initial test suite with GTX 980 (PC specs below)*:

 

Test with VR and specified VR test settings:  Frames: 3461 - Time: 60000ms - Avg: 57.683 - Min: 35 - Max: 63

 

Test with VR (1.7 SS) and my usual monitor settings**:  Frames: 3331 - Time: 60000ms - Avg: 55.517 - Min: 30 - Max: 62

 

Test with monitor and my usual monitor settings**:  Frames: 7014 - Time: 60000ms - Avg: 116.900 - Min: 71 - Max: 153

 

*old PC specs:  Intel Core i7-4770K @3.5 GHz; ASUS ROG Maximus VI Hero MB; 16GB RAM (DDR3 1600MHz)

EVGA GTX 980, 4 GB; Realtek ROG SupremeFX audio
Windows 10 Pro 64-bit

 

**usual monitor settings (differences from specified VR test settings) = 1920x1080; Ultra; mirrors = complex; distant landscape detail = 4; grass quality = distant; Antialiasing = 4.

 

 

Second test suite, after upgrade to GTX 1080TI (same specs as above except for Video card):

 

Test with VR and specified VR test settings:  Frames: 4554 - Time: 60000ms - Avg: 75.900 - Min: 61 - Max: 94

 

Test with VR (1.7 SS) and my usual monitor settings: Frames: 4017 - Time: 60000ms - Avg: 66.950 - Min: 54 - Max: 81

 

Test with monitor and my usual monitor settings:  Frames: 7490 - Time: 60000ms - Avg: 124.833 - Min: 97 - Max: 160

 

 

Third test suite, after upgrade to new PC (specs below*):

 

Test with VR and specified VR test settings:  Frames: 5029 - Time: 60000ms - Avg: 83.817 - Min: 50 - Max: 91

 

Test with VR (1.7 SS) and my usual monitor settings: Frames: 4938 - Time: 60000ms - Avg: 82.300 - Min: 52 - Max: 91

 

Test with monitor and my usual monitor settings:  Frames: 9087 - Time: 60000ms - Avg: 151.450 - Min: 119 - Max: 192

 

*new PC specs:  Intel Core i7-7700K @4.2 GHz; ASUS ROG Maximus IX Code MB; 16GB RAM (DDR4 3200MHz)

EVGA GTX 1080 TI SC2, 11 GB; Realtek ROG SupremeFX audio
Windows 10 Pro 64-bit
 
I'm puzzled by some comment made a few posts above that the max FPS shouldn't be more than 90 on the Rift, so I'm not sure what FRAPS is actually measuring here, but the info above is copy/pasted out of the FRAPS results files.
 
[edit]  latest test (with 1st post VR test settings) re-run and updated above to reflect Rift headset definitely activated and SSAO off per latest test parameters
Edited by TG-55Panthercules

Share this post


Link to post
Share on other sites

@Panthercules

 

Wow, that's a very powerful rig you have now!

 

What Chiliwili meant above is that in VR you are capped at 91fps. There is no need to go higher than what the VR hmd screen can handle, so that is the reason for the cap.

 

When benchmarking VR make sure that your rift is activated. Meaning if you have it on a table you should see that the lenses are on and showing the same image you have on your monitor. Chiliwili suspects that since you got results above 91fps your rift might not have been active when you benchmarked.

 

By placing something in front of the light sensitive sensor inside the hmd you can have it activated without wearing it.

 

I have also gotten results above 91 max fps. In my case the rift was activated but obviously something else was wrong, restarting the computer fixed this.

Share this post


Link to post
Share on other sites

 

 

I'm puzzled by some comment made a few posts above that the max FPS shouldn't be more than 90 on the Rift, so I'm not sure what FRAPS is actually measuring here, but the info above is copy/pasted out of the FRAPS results files.

 

Many thanks for sharing your new performance in your new PC, you have done a good upgrade.

 

FRAPS report the same fps than reported by BoS ingame counter, it can be what ever (no limit in the Fraps side). When you are playing with BoS in VR you will see that the counter never exceed 91.

 

When Fraps report more than 91 for a VR test is because something in the test is wrongly done. You see correctly the Balapan track in the monitor (the VR right eye view) but if you can check waht is visualized in the rift you will see that the Balapan track is not displayed (instead you might see the SteamVR sensors scene)

 

I see that the VR test with the procedure settings is wrong (Max higher than 91) but the VR test with your usual setting is correctly done (Max=91)

 

If you still keep the 980, it would be interesting for many people to see how it performs in your new rig.

 

BTW, if you have a good CPU cooler you still will improve even more your results if you OC your CPU (radek is running at 4.8)


 

 

Test with VR and specified VR test settings:  Frames: 5947 - Time: 60000ms - Avg: 99.117 - Min: 81 - Max: 123

 

If you repeat this test (since Max should be 91 or less) make sure you put SSAO off.

Two days ago, we changed that setting in the procedure.

Share this post


Link to post
Share on other sites

Feel I have to give out a warning here for VR users.

 

My earlier benchmark with an overclocked i7 7700 and a gtx1070 according to settings in the first post, had an average of 72.244 fps.

 

This did not match chiliwilis result at an average of 81.12 fps, as he is running the same 1070gpu but an older CPU.

 

Took some time finding the cause. But finally remembered I had installed a 4K skinpack I had found on this forum. This skinpack also contained large resolution bump maps replacing the standard ones. I suspect those were the culprit as the Balapan track would not use any regular custom skins.

 

Re-installed il2 and benchmarked again. New result is 85.85 fps average. A 13fps difference on my rig. While a more powerful Gpu might handle that skinpack better, 13fps with a gtx1070 could be the difference between screen tearing and silky smooth 90fps. With a card slower than the gtx1070 the difference would be even bigger.

 

Results in the spreadsheet updated.

 

https://docs.google.com/spreadsheets/d/1gJmnz_nVxI6_dG_UYNCCpZVK2-f8NBy-y1gia77Hu_k/edit?usp=sharing

Share this post


Link to post
Share on other sites

OK - I'll re-run the tests, making sure that the headset is actually on this time.  Unfortunately, I can only do it with my current machine, so if there were issues with the previous test runs on the earlier machine configs those results comparisons won't be worth much :(

 

I realized yesterday that the SSAO aspect of the test procedures had changed since I ran my earlier tests, so I ran my latest test set with SSAO both on and off, and didn't see a significant difference (about 1 FPS, within margin of error in iterative runs), but I'll make sure to run this next set with it off.

 

[edit]  Re-ran the test with 1st post settings and my post with results (#103) above has been updated as noted above.  Pretty decent improvement from my old rig to my new one:

 

VR test settings:  Avg = + 26 FPS  (+ 45%);  Min = + 15 FPS  (+ 42%);  Max = + 28 FPS  (+ 44%)

 

My VR/monitor settings:  Avg = +  27 FPS  (+ 48%);  Min = + 22 FPS  (+ 73%);  Max = + 29 FPS  (+ 47%)

 

 

So, I'm happier about the upgrade than I thought I was based on the first try at the testing.  Looking forward to how it works out while actually playing the game.

Edited by TG-55Panthercules

Share this post


Link to post
Share on other sites

Thanks Panthercules for re-run the test with your new rig. I have updated your test in the spreadsheet. You have a now a very good performance and a lot of room to go further with the SS if required.

 

If I go to 2.25 SS (equivalent to 1.5 PD) my avg fps drop from 81.12 to 70.7. Since you have a 1080Ti, I think that your system will be less impacted by SS.

 

You will know that the 7700K can support well some OC, so if in the future you need more power from the CPU and you have a good CPU cooler you can safely OC to 4.6 or 4.8.

 

Enjoy VR!

Share this post


Link to post
Share on other sites

Panthercules. Out of curiosity, have you enabled xmp yet or manually configured your memory to run at 3200?

Share this post


Link to post
Share on other sites

Panthercules. Out of curiosity, have you enabled xmp yet or manually configured your memory to run at 3200?

 

Could not find an XMP setting as such in the BIOS (way too many options - rather confusing) but I did find a place to manually set it to 3200 (and it's now being reported at 3200 instead of 2133 or whatever it was before, so I'm pretty sure it's working).  Haven't messed with OC'ing my CPU yet - wanted to get some solid benchmarks run at stock 4.2 before I started testing higher settings there.  Also wanted to get a good feel for temps and noise at stock before doing that as well (so far, so good - nice and quiet and staying cool so far).  

 

Just did a QMB (in VR but using my normal monitor settings with things cranked up) with a few IL-2s facing a few Ju-52s over winter Stalingrad - was looking really awesome before one of the Ju-52 gunners took out my engine and I glided to a reasonably safe landing on the frozen river (nosed over a bit but survived largely intact).  Had Fraps on during the mission and it recorded Avg: 87.854 - Min: 46 - Max: 91 - not scientific at all, but a pretty good indication I think that things are working well in a game setting with this rig.  Out of time for any more testing today, but am looking forward to some more serious flying time tomorrow.  I think I'm really gonna enjoy this new toy and will finally get into BoS in a bigger way soon.

Edited by TG-55Panthercules

Share this post


Link to post
Share on other sites

 

Just did a QMB (in VR but using my normal monitor settings with things cranked up) with a few IL-2s facing a few Ju-52s over winter Stalingrad - was looking really awesome before one of the Ju-52 gunners took out my engine and I glided to a reasonably safe landing on the frozen river (nosed over a bit but survived largely intact).  Had Fraps on during the mission and it recorded Avg: 87.854 - Min: 46 - Max: 91 - not scientific at all, but a pretty good indication I think that things are working well in a game setting with this rig.  Out of time for any more testing today, but am looking forward to some more serious flying time tomorrow.  I think I'm really gonna enjoy this new toy and will finally get into BoS in a bigger way soon.

 

:good: 

 

I think you are golden, have fun !!

 

I too have found that since VR, I have really gotten into BoS like I never had before.

Share this post


Link to post
Share on other sites

IMHO this is a very useful thread. Much to useful to be limited only by VR measurements. Guys who feed the spreadsheet please consider moving this thread out of VR forums to general hardware discussion and adding instructions for monitor users. We will gather more statistics then.

Share this post


Link to post
Share on other sites

@Panthercules

 

I understand you want to fly some now with the excellent performance you already have.

 

When you feel like going back to tweaking, the 7700k is remarkably easy to OC. Only three clicks including save&exit. This is not related to benchmarking, only want to give this advice.

 

In the first screen you see upon entering your "uefi bios ez mode", on the left in the middle, right above the spinning fans, will be a drop down menu called x.m.p.

 

All that's needed is to switch it from disabled to your 3200 xmp profile. Once you do this you get a pop up asking if you want to bump your baseclock as well. If you select yes, you get an instant CPU overclock to 4.5ghz. Save and exit.

 

That's it! You'l be running your memory at optimal factory settings and your CPU at a very safe 4.5ghz. If you want to go further good cooling is needed and some more work.

 

Be sure to reset all you already did to default first. If I understood you correctly, what you found and already did is actually harder than just enabling your xmp profile.

Share this post


Link to post
Share on other sites

 

 

IMHO this is a very useful thread. Much to useful to be limited only by VR measurements. Guys who feed the spreadsheet please consider moving this thread out of VR forums to general hardware discussion and adding instructions for monitor users. We will gather more statistics then.

 

Initially the balapan tests in the russian forum was recording both, performance in VR and monitor. But the focus was VR, they put it monitor just for comparison.

 

I also think it would be useful to add monitor performance, but as a complementary information to the VR test. So, in this way, people without VR device can run the test in monitor and figure out what performance they will achieve with VR.

 

I only fly in VR, (I think radek as well), so our interest with this thread is VR performance. I also understand that many people (most of them) still fly in monitor (because they are waiting for 2nd generation or other reasons), but in the next years there will be a gradual migration of simmers to VR. There are also many configurations of resolutions (1080p, 2K, 4K), single-triple monitors, singleGPU-SLI, etc. So it is difficult to cover all of them in a benchmark.

 

I would propose to define the settings for the monitor the same than the VR-test settings, but with resolution at FullHD (1080p) and AAx4. 

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.

 

For monitor only simmers it should help with comparing performance, know what upgrades would make the biggest difference and of course also help those on older rigs to determine if theirs can handle this sim at all.

 

Majority probably have a 1080p capable monitor today. So it's a good resolution. My proposal would be to have terrain distance at a maximum of 2x. This so we don't exclude those running older Gpu's with less memory.

Share this post


Link to post
Share on other sites

@Panthercules

 

I understand you want to fly some now with the excellent performance you already have.

 

When you feel like going back to tweaking, the 7700k is remarkably easy to OC. Only three clicks including save&exit. This is not related to benchmarking, only want to give this advice.

 

In the first screen you see upon entering your "uefi bios ez mode", on the left in the middle, right above the spinning fans, will be a drop down menu called x.m.p.

 

All that's needed is to switch it from disabled to your 3200 xmp profile. Once you do this you get a pop up asking if you want to bump your baseclock as well. If you select yes, you get an instant CPU overclock to 4.5ghz. Save and exit.

 

That's it! You'l be running your memory at optimal factory settings and your CPU at a very safe 4.5ghz. If you want to go further good cooling is needed and some more work.

 

Be sure to reset all you already did to default first. If I understood you correctly, what you found and already did is actually harder than just enabling your xmp profile.

 

Thanks for the tips - it took a couple of tries to make it stick (doing the xmp thing didn't seem to work - boosted the ram but not the CPU, despite telling it yes on the baseclock thing), but I ran through the EZ wizard and that seemed to do the trick).

 

Anyway, re-ran the Passmark tests and saw the following (going from 4.2 to 4.5):

 

Passmark rating:  6622.1 to 7056.7  (6.5% increase)

CPU Mark:  13044.5 to 14052.9  (7.7% increase)

3D Graphics Mark:  17854.9 to 18628.0 (4.3% increase)

 

Will re-run the VR benchmark next and see what happens.

 

[EDIT] - OK, now I'm very confused - test with the VR benchmark settings went down:

 

Avg: 81.183 - Min: 43 - Max: 91  (had been Avg: 83.817 - Min: 50 - Max: 91)

 

Gonna have to recheck all my settings and make sure something didn't get changed by mistake somewhere along the way.  

Edited by TG-55Panthercules

Share this post


Link to post
Share on other sites

Regarding il2 benchmark. had that happen to me, I my case, asw was not properly off :)

 

Glad it worked. Never had any trouble getting the xmp CPU thing stick.

Share this post


Link to post
Share on other sites

Well, no idea what's going on, but after resetting the MB to defaults and re-doing the memory to 3200 the manual way I did it originally, the VR benchmark test results are back up to where they were before, higher than when I tried OC'ing the CPU to 4.5.

 

Think I'll quit while I'm ahead, and just go with what I've got - saw some other posts around here somewhere that said they were seeing relatively good gains up to 4.2 but relatively little above 4.2, so seems to be little point in chasing those last few increments anyway.

Edited by TG-55Panthercules

Share this post


Link to post
Share on other sites

Very strange what happen with the OC test. I reported my OC´s test in monitor and VR in the thread https://forum.il2sturmovik.com/topic/29881-overclocking-4790k-better-bos-performance

There always an increase even around 4.6.

 

To OC my CPU I didn´t do it through BIOS. I used a simpler software tool called MSI Command Center. Pretty simple for moderate OC.

 

There are other software tools for other MoBo manufacturers, For ASUS there is this one: http://asus-turbov-evo.en.lo4d.com/

But not sure if it will work with your MoBo.

Share this post


Link to post
Share on other sites

Well, something weird is still happening - although I reset everything to defaults in BIOS and can't see any indication there that the CPU is still OC'd (may just not know where to look - way too many settings in there), when I run CPU-Z it shows the following (looks to me like it's got the CPU at 4.5):

 

post-12147-0-90630500-1502404419_thumb.jpg

 

Not sure if it's still OC'd or if it's kicked into "turbo mode" (it says target for turbo mode is 4.5, but it's not clear (to me) from the BIOS when turbo mode will or won't be active).

 

Bottom line, I'm not sure if my latest set of results is with the CPU at 4.2 or 4.5.  Is there some way to tell what the CPU is actually operating at during the time the Fraps test is running?

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...