Jump to content
chiliwili69

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

Recommended Posts

2 hours ago, 19//SAG said:

Here are my results after years thinking i would be upgrading my pc soon (turns out intel or nvidia don't really want to push tech just yet)

 

Motherboard
Asrock Z87 Extreme6

 

Processor
I5 4670K @ 4.3GHZ

 

RAM
Kingston HyperX Fury 8x2GB 1600

 

GFX
MSI GTX 1070 8GB VRAM

 

Storage (for the game)
Samsung SSD 120gb
 

VR Headset

Oculus Rift CV1

 

CPU MARK

9299

 

CPU single threaded

2567

 

Frames: 5373

Time: 120000ms

Avg: 44.775 - Min: 39 - Max: 50

 

IDK if upgrading to an 8700k and a 1080ti or 2080ti would be worth it for me, i dont really mind the slow framerates (even when landing at a very busy airfield when i get 25ish fps) What do you guys think?

 

Did you disable ASW while doing the test?

Share this post


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

 

Did you disable ASW while doing the test?

 

I did :( 

 

Share this post


Link to post
Share on other sites

I joust upgraded my computer from  i5 7400 3,00GHz 8 gb ram GTX 1070 to i5 7400 3,00Ghz 16Gb Ram Msi 1080Ti 

What should I expect regarding fps and graphics settings along with super sampling? If this is the wrong thread please show me the way to the right one.

Share this post


Link to post
Share on other sites
On 1/10/2019 at 5:53 PM, 19//SAG said:

Avg: 44.775 - Min: 39 - Max: 50

 

Thank you for running the test with your i5-4670K. It is the first model of this CPU tested in the table.

According to your Singlethread value 2567 you should achieve a bit more.

There could be 4 reasons for that:

- The 1070 is being bottleneck. You can check that by looking the load of your GPU during the track run. You can do that with MSI afterburner free tool.

- The RAM 1600 is affecting the performance

- The i5-4670K suffer the same problem than the i5-4690K, it was the only CPU that didn´t correlate well with the singlethread vs Avg-fps

- Something is not well configured in your system or throttling the CPU during run. Look this checklist: https://forum.il2sturmovik.com/topic/34107-items-to-review-if-you-have-low-performance-in-il-2-vr-test/

 

But if you are fine with your framerates, then there is no reason to upgrade. But if one day you want to upgrade I would upgrade first the CPU and keep your GPU.

 

  • Thanks 1

Share this post


Link to post
Share on other sites

If upgrading, I would consider new MB, Ram and CPU rather than just the CPU.

Edited by dburne
  • Thanks 1

Share this post


Link to post
Share on other sites
On 1/11/2019 at 10:02 AM, WhiteMoose said:

I joust upgraded my computer from  i5 7400 3,00GHz 8 gb ram GTX 1070 to i5 7400 3,00Ghz 16Gb Ram Msi 1080Ti 

What should I expect regarding fps and graphics settings along with super sampling? If this is the wrong thread please show me the way to the right one.

 

Unfortunately I think that CPU is going to hold you back. The GPU will be able to do increased supersampling, so the picture should look nice and crisp. I would start with SteamVR 150% supersample and 'balanced' in-game settings, and see how it goes. A lot of people are fine flying 45 FPS ASW, if that works for you then you can increase to 'high' in game settings.

Share this post


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

 

Unfortunately I think that CPU is going to hold you back. The GPU will be able to do increased supersampling, so the picture should look nice and crisp. I would start with SteamVR 150% supersample and 'balanced' in-game settings, and see how it goes. A lot of people are fine flying 45 FPS ASW, if that works for you then you can increase to 'high' in game settings.

What processor would you recommend to get things rollin?

Share this post


Link to post
Share on other sites
13 minutes ago, WhiteMoose said:

What processor would you recommend to get things rollin?

 

I'm personally thinking on a 8700K with 16gb of 3000ram, but of course, it depends on your budget.

Share this post


Link to post
Share on other sites

Salut, Servus,

core i5 9600k@5040.3MHz (1,39V set in BIOS); MSI MPG Z390 G+; Samsung 860 pro; Corsair LPX DDR4 3200MHz C16 (XMP set in BIOS); GTX 1080Ti

 

SS 1.5 - Graphic settings set as instructed

2019-02-03 13:20:36 - Il-2
Frames: 9921 - Time: 120000ms - Avg: 82.675 - Min: 43 - Max: 91

LB

 

 

 

Edited by Laurentlb33

Share this post


Link to post
Share on other sites
On 2/3/2019 at 12:53 PM, Laurentlb33 said:

core i5 9600k

 

Thank you LB for giving our first test with the 9600K at 5.0GHz.

I assume you use the Rift, right?

Also, could you run the passmark test and provide the single-thread performance as explained in the instructions?

I also assume you did it with version 3.009.  It will be our last run of that version.

 

Have anybody tried the v3.010 with this benchmark?  I have sold my Rift and can not compare.

 

Note: The test with Pimax5K+ is still to be done with this benchmark. I really don´t know what SS should I use here just to compare performance with other headset in a fair way.

Share this post


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

I really don´t know what SS should I use here just to compare performance with other headset in a fair way


1.0 and 'normal' fov. If I may make a suggestion.

Not a fair performance comparision with gen1 hmd's but I don't see the need for that. Most would buy the pimax for the improved image clarity/fov and so also expect a drop in performance.  For me performance comparisons between pimax users and what to expect at 1.0 ss with any given hardware is plenty more interesting.

Share this post


Link to post
Share on other sites

I suppose that with 1.0 you mean Pittol at 1.0 and SteamVR at 100%.  Right?

If this is the case then the clear bottleneck is GPU. 

 

My idea was to use 66% (or something like that) which also gives good visuals (superior to Rift) but also a more decent performance.

I am also thinking about other display frequencies (60, 72, 90) and other FFR modes.

Share this post


Link to post
Share on other sites

Well, I have just run the first benchmark over the Pimax5K+ and IL-2 version 3.010

 

I have used the Normal FOV, Pitool Rendering Quality=1.0 and SteamVR SS set at 26%. This gives 6.59 million pixels which is exactly the same amount of pixels than the Rift at 150% SteamVR SS. I still didn´t upgrade to the new Pitool with freq changes, so it is 90Hz for the display.

 

The result for my CPU at 4.8GHz is

Frames: 9674, Avg: 80.617, Min:62, Max:91

 

I chose 26% to be coherent with the rules I applied to the VivePro and Odyssey+.

I don´t use 26% at game (I have not tried how it looks like), normally I use 66% and SteamVR recommendation is 55% for my 1080Ti.

 

But what everybody will want to know is how the performance change when the SS is increased, so I conducted a series of tests from 26% to 150% SS in SteamVR.

The results are:

1265448727_Pimax5KandSStests.thumb.png.f0bc60158a090f226db60c1650c154b6.png

 

or better in a graph:

1989997655_Pimax5KandSStestsgraph.png.d79129be34cf96337462240c47ab27ab.png

 

this graph is in the "SS per device" online spreadsheet of the benchmark.

 

So, WELCOME again to the GPU bottleneck AGE!!

 

Now it is a tradeoff of visual quality versus fps. Of course, good overclocked CPUs is a must but not enough !  :huh:

 

Any Pimax5K+ user is welcome to run the test as well, specially the ones with 20xx cards for the FFR.

 

 

Edited by chiliwili69
  • Like 2
  • Thanks 1

Share this post


Link to post
Share on other sites

So we need a 4080ti :) This is great info chili and based on this I'm getting me a new card and a pimax. Pleasantly surprised your avg fps at 26% steam ss is so close to your rift performance also. Expected pimax not to reach quite all the way so soon considering they haven't had those years of optimization.

Share this post


Link to post
Share on other sites

Honestly I'm really curious to see how the Radeon VII performs in IL-2. Not that I can upgrade to it. Currently using a 1080, can't justify ~1k cad for 20% at best... same reason I can't go to the 2080 series... just isn't worth it.

Edited by R3animate

Share this post


Link to post
Share on other sites

Hey guys, super new to IL2, been trying to get my VR up and running and finally have a score worth posting after some trouble shooting!

Game Version 

3.010b

 

Motherboard
ASUS Z270E STRIX

 

Processor
I7 7700k @ 4.7GHz

 

RAM
G.Skill Trident Z 32GB DDR4 @ 3200MHz

 

GFX
NVIDIA FE 1080ti

 

Storage (for the game)
Samsung 960 EVO m.2 NVMe
 

VR Headset

Oculus Rift

 

CPU MARK

10675

 

CPU single threaded

2605

 

Frames: 8633

Time: 120000ms

Avg: 71.942 - Min: 43 - Max: 91

 

I've just put a custom watercooling loop in and have been having some issues with CPU temp. I know the 7700k is notoriously bad for this, but I'll be draining it this weekend and reapplying the thermal paste in an attempt to get a higher clock.

 

Thanks!

 

EDIT: Decided to disable Hyperthreading and turn up the overclock a bit. Running it with Hyperthreading disabled and a clock of 4.9GHz resulted in the following: (Note: My average actually went down slightly while my min and max both jumped up.)

 

CPU MARK

11406

 

CPU single threaded

2837

 

Frames: 8540

Time: 120000ms

Avg: 71.167 - Min: 49 - Max: 94

Edited by iNcHeS

Share this post


Link to post
Share on other sites

Here is my new test with the Pimax 5K.
This test proves that the formula does not apply:
More FPS equals = is smoother

 

After trying a lot, I wanted the best result between sharp and smooth.

 

I9-9900K @ 5.1GHz WaterCool Max Temp: 65 ° C
RTX 2080TI Max Overclocked Core 2010 MHz, Memory 7000MHz WaterCool

 

Bios = everything on Auto

 

Pimax and Steam:

Pitool version ... 91, factor 1.75
Steam VR APP = 100%, 3162x2701- Video: 22% 3162x2701
Motion Smoothing OFF

3DMigoto Mod = ON

 

IL2 Setting:
Main Settings High
Shadow = Max
Mirror = Off
Long textures = x4
View = 70Km
Ani. = Blurred
Grass = Normal
Clouds = High
FDA = Full
AA = x2
Gamma = 0.8
Focus = on
4K Textures On

 

2019-02-09 14:06:34 - Il-2
Frames: 5489 - Time: 120000ms - Avg: 45.742 - Min: 40 - Max: 50

 

So I have about 5000 FPS less to my previous measurement with all the settings from the first post.

When I play with the current setting, everything is absolutely smooth and the tracking with two ViveBoxes is absolutely fantastic.
No matter if online or offline.
No matter if many clouds or little.
No matter if busy or not.

I feel it as smooth as with 5000FPS more.

Edited by I./JG68_Sperber
  • Like 1

Share this post


Link to post
Share on other sites
On 2/5/2019 at 9:22 AM, chiliwili69 said:

 

Thank you LB for giving our first test with the 9600K at 5.0GHz.

I assume you use the Rift, right?

Also, could you run the passmark test and provide the single-thread performance as explained in the instructions?

I also assume you did it with version 3.009.  It will be our last run of that version.

 

Have anybody tried the v3.010 with this benchmark?  I have sold my Rift and can not compare.

 

Note: The test with Pimax5K+ is still to be done with this benchmark. I really don´t know what SS should I use here just to compare performance with other headset in a fair way.

I am actually using the Rift. Here is the passmark result. The test was done in 3.009

 

Capture.PNG.2860bdecaf9dc3a6e4c3becc88053abe.PNG

Share this post


Link to post
Share on other sites
On 2/8/2019 at 12:37 PM, R3animate said:

Honestly I'm really curious to see how the Radeon VII performs in IL-2. Not that I can upgrade to it. Currently using a 1080, can't justify ~1k cad for 20% at best... same reason I can't go to the 2080 series... just isn't worth it.

The only card to buy for that price is a 2080ti. Nothing less. The 2080ti is by far the fastest card out right now. Nothing comes close. It's like 35-40% faster than a 1080ti or 2080. I should sell my 1080ti and 1070 I have and try and get one.

Edited by 392FS_Jred

Share this post


Link to post
Share on other sites
On 2/9/2019 at 6:44 AM, iNcHeS said:

EDIT: Decided to disable Hyperthreading and turn up the overclock a bit. Running it with Hyperthreading disabled and a clock of 4.9GHz resulted in the following: (Note: My average actually went down slightly while my min and max both jumped up.)

 

CPU MARK

11406

 

CPU single threaded

2837

 

Frames: 8540

Time: 120000ms

Avg: 71.167 - Min: 49 - Max: 94

 

Welcome to VR club and thank you for posting your tests.

 

In your first test you use OC in CPU with 4.7GHz and achieve STMark of only 2605. This is not aligned with previous tests of 7700K. Look at the  "STMark vs OC" tab in the spreadsheet and you will see that for the 7700K for 4.7 the expected STMark value should be around 2750, not 2605.

Maybe you were running the STMark with different OC (4.5 perhaps) or maybe your CPU is throttling while you run the CPU Passmark.

 

In you second test, you achieve a Max of 94. This is not possible. Typically the maximun Max is 90 or 91, but not 94.

But in this case the STMark you achieve (2837) is very much aligned with the STMark expected for a 7700K at 4.9 GHz.

 

In theory your 1080Ti should not be fully loaded with this test. To check that you can run MSI Afterburner during the IL-2 test and use their trendlines to visualize CPU freq., CPU temp, GPU freq, GPU load and GPU temp.

If GPU is at 100% verify your Supersampling values (what do you use SteamVR or OTT?)

 

Share this post


Link to post
Share on other sites
On 2/9/2019 at 2:49 PM, I./JG68_Sperber said:

Here is my new test with the Pimax 5K.
This test proves that the formula does not apply:
More FPS equals = is smoother

 

After trying a lot, I wanted the best result between sharp and smooth.

 

I9-9900K @ 5.1GHz WaterCool Max Temp: 65 ° C
RTX 2080TI Max Overclocked Core 2010 MHz, Memory 7000MHz WaterCool

 

Bios = everything on Auto

 

Pimax and Steam:

Pitool version ... 91, factor 1.75
Steam VR APP = 100%, 3162x2701- Video: 22% 3162x2701
Motion Smoothing OFF

3DMigoto Mod = ON

 

Thank you for posting your personal results with you superbeast machine and the Pimax5K. 

Achieving 5.1GHz in this 9900K is a great achievement.

 

Please, if you could, post also the Passmark test as indicated in the instructions (Passmark is a free non-invasive software).

 

Since it is not using the same settings that the test I didn´t put your results in the main page, but in the "SS tests".

 

I see you use Pitool RQ=1.75 and SteamVR at 22%. This gives a multiplying factor of 1.75x1.75x0.22= 0.67, so your test is equivalent to running RQ=1 and SteamVR SS at 67% (or 66%).

250374559_pimax5KRQ1_75.jpg.ada785656bc8ea58b92f6e11089d51a9.jpg

 

There is no reason to touch pitool RQ, I have it always at 1.0. The graphical effect is the same than touching SS, I demonstrated that in this post.

Someone say that using higher than 1.0 in RQ and reduce SteamVRSS for the same pixels gives better performance, but he didn´t demonstrated that. So, I don´t think this influence performance at all. (but still to be proved).

 

For for the shake of comparing your performance with others, you could run the test exactly at the same setting that instructed in the post and share your results.

 

Running at 45fps could be very smooth for many people, it is very subjective. I also have to say (it is also subjective) that I notice less the lack of being at 90fps in the Pimax5K.

 

 

Share this post


Link to post
Share on other sites
On 2/15/2019 at 5:46 PM, I./JG68_Sperber said:

Hi chiliwili69,

 

Pimax 1 and Steam 66% i have 5 FPS more ;)

 

Here my Passmark 5,1 GHz

 

pass.jpg

Try to aim for a vertical res of a factor of 1440

Maybe PiTool 1.75 or 2

SteamVR Video 20%

App setting then as close to 2880 as possible which would be a SS factor of x2 from the 5K native Res.

 

You could also just try the native resolution on the Vertical of 1440 in SteamVR App setting and see how that looks on the 5K.  It doesn't look good on the 8K but 2160 does.

Share this post


Link to post
Share on other sites
On 2/15/2019 at 7:25 PM, chiliwili69 said:

 

Welcome to VR club and thank you for posting your tests.

 

In your first test you use OC in CPU with 4.7GHz and achieve STMark of only 2605. This is not aligned with previous tests of 7700K. Look at the  "STMark vs OC" tab in the spreadsheet and you will see that for the 7700K for 4.7 the expected STMark value should be around 2750, not 2605.

Maybe you were running the STMark with different OC (4.5 perhaps) or maybe your CPU is throttling while you run the CPU Passmark.

 

In you second test, you achieve a Max of 94. This is not possible. Typically the maximun Max is 90 or 91, but not 94.

But in this case the STMark you achieve (2837) is very much aligned with the STMark expected for a 7700K at 4.9 GHz.

 

In theory your 1080Ti should not be fully loaded with this test. To check that you can run MSI Afterburner during the IL-2 test and use their trendlines to visualize CPU freq., CPU temp, GPU freq, GPU load and GPU temp.

If GPU is at 100% verify your Supersampling values (what do you use SteamVR or OTT?)

 

 

Thanks for the feedback - I was having a few issues initially getting anywhere near those scores. STMark was definitely run at 4.7 - checked using Argus monitor before running the tests.

 

I did think it was weird at the time seeing 94 as no one else had gotten above 91, any idea what could have caused it? I've got MSI Afterburner installed now. In regards to the Supersampling values, that was one thing I thought was quite strange. Setting my (SteamVR) SS to 150% results in a different eye render than what is displayed on the instruction page. At 150%, mine says a render of 1626x1940 per eye. Is that normal?

 

Cheers

Share this post


Link to post
Share on other sites
8 hours ago, iNcHeS said:

checked using Argus monitor before running the tests.

 

Checking the freq before the test doesn´t guarantee that you maintain that freq during the test, since if you have high temp, or high amps,or high power then your CPU could throttle (reduce the freq). That why is good to run MSI Afterbuner trendlines during the test, so you can see freq is maintained.

 

The 94 could be due some short of anomaly during the test. Try to repeat the test to see if it is again 94 the max fps.

 

The 1626x1940 is also strange. Perhaps Oculus has upgraded the software with a bit lower internal SS. I have not a Rift to measure again. Can someone else verify this?

Another reason could be that you have a small subsampling somewhere else, for example having in OTT a 0.99 instead of a 1.0.

Or 98% in the SteamVR App section.

Share this post


Link to post
Share on other sites

Today i run the same test with HT disabled and CPU overcklocked to 4,9ghz - Test results were the same from the previous 4,3ghz test... any idea why i din't got any gain?

 

The cpu is capping the gpu, but increasing the oc didn't get results...

Tempertures were stable in the rage of 70°.... so i doubt ther's any throttling...

 

 

 

anyway i'm tempted to upgrade to a Ryzen 2700x

Asus Crosshair VII hero

and 32gb (damn DCS) of 3200 cas 14 ram... (samsung B die)

 

Comments on this build??

 

Also I checked the spreadsheet and didn't find any Ryzen yet....! Could be a first one...

Share this post


Link to post
Share on other sites

There is a Ryzen 2700x in table: 46 avg FPS, not very impressive. Try to switch tabs in the bottom of spreadsheet page

Share this post


Link to post
Share on other sites
14 hours ago, EAF51_Jimmi said:

Today i run the same test with HT disabled and CPU overcklocked to 4,9ghz - Test results were the same from the previous 4,3ghz test... any idea why i din't got any gain?

 

The cpu is capping the gpu, but increasing the oc didn't get results...

Tempertures were stable in the rage of 70°.... so i doubt ther's any throttling...

 

 

 

anyway i'm tempted to upgrade to a Ryzen 2700x

Asus Crosshair VII hero

and 32gb (damn DCS) of 3200 cas 14 ram... (samsung B die)

 

Comments on this build??

 

Also I checked the spreadsheet and didn't find any Ryzen yet....! Could be a first one...

Please do not handicap yourself with inferior products such as a ryzen series cpu. If you are deadset on supporting a company that hasn't produced anything competitive in nearly a decade then wait for the ryzen 3 to show up in the hopes it may finally reach parity with intel.

Share this post


Link to post
Share on other sites
Posted (edited)

Ok, here goes.

 

Laptop - Gigabyte Aero 15X

i7700HQ, 32Gb DDR4 2666 Ram, NVME 512Gb drive, GTX 1070 Max-Q

eGPU Gigabyte RTX 2080 Gaming OC with Razor Core housing

 

Tested with the Pimax 5K setting recommended, PiTool at 1, SteamVR Video 26%, SteamVR App 100% giving a target resolution of 2260 x 1991

 

image.png.35a094c94ef72ca49f336c8e64c5b398.png

 

Test run on a Pimax 8K HMD

 

First run on the RTX 2080

Frames Avg 44.417, Min 30, High 55

 

Second run on the GTX 1070 Max-Q

Frames Avg 30.225, Min 33, Max 44

 

Pre eGPU I would run it on Low, with Terrain render at 4x, Distance at 100, Clouds and Shadows High, Sharpening, SSOA, HDR and 4K textures.  I can't remember the original Target Render Resolution as there has been so much experimenting since.

 

With the eGPU I run it on Balanced the same options as Pre eGPU.  The Test track and settings are an odd bunch.  Target Render Settings I am testing at the moment are 3469 x 3055 and 2453 x 2160.  I think the later is better not just in frames but also with Nvidia Control Panel AntiAlaising Transparency Setting set to SS x2 or x4 - I get a pretty decent view in headset which makes spotting good both air and ground targets and decent frames.  Normal AntiAliasing I leave off.

 

In test maybe 30% better with eGPU but it also scales well with res and my Laptop doesn't cook itself.

image.thumb.png.71abc89737cdf27b67580eb260741cb9.png

4 hours ago, JonRedcorn said:

Please do not handicap yourself with inferior products such as a ryzen series cpu. If you are deadset on supporting a company that hasn't produced anything competitive in nearly a decade then wait for the ryzen 3 to show up in the hopes it may finally reach parity with intel.

Not that they are inferior - in many cases they are superior but at the moment in Single Threaded Processes - Intel wins.  Unfortunately for us, Il2 relies strongly on single threaded performance at this point in time. 

 

Ryzen 3 might be worth a look at when it is released but wait until release and reviews are in before jumping on it.  If you are wanting a CPU for a Server or a Workstation PC - then AMD all the way.  Not to mention you'd be supporting Canuks instead of Kosha Nostra.

Edited by blitze
Forgot the to mention the Headset settings

Share this post


Link to post
Share on other sites
Posted (edited)
18 hours ago, EAF51_Jimmi said:

Today i run the same test with HT disabled and CPU overcklocked to 4,9ghz - Test results were the same from the previous 4,3ghz test... any idea why i din't got any gain?

 

The cpu is capping the gpu, but increasing the oc didn't get results...

Tempertures were stable in the rage of 70°.... so i doubt ther's any throttling...

 

 

 

anyway i'm tempted to upgrade to a Ryzen 2700x

Asus Crosshair VII hero

and 32gb (damn DCS) of 3200 cas 14 ram... (samsung B die)

 

Comments on this build??

 

Also I checked the spreadsheet and didn't find any Ryzen yet....! Could be a first one...

You may want to check whether you have forgotten to set the AVX offset to 0.

 

P.S. As said before, Ryzen 2700x would be a downgrade for IL-2 performance. The game heavily depends upon single core clock speed. A quick RAM speed (3200MHz) has shown to be quite beneficial and to boost fps considerably. That's the right direction 👍

On 2/3/2019 at 12:53 PM, Laurentlb33 said:

Salut, Servus,

core i5 9600k@5040.3MHz (1,39V set in BIOS); MSI MPG Z390 G+; Samsung 860 pro; Corsair LPX DDR4 3200MHz C16 (XMP set in BIOS); GTX 1080Ti

 

SS 1.5 - Graphic settings set as instructed

2019-02-03 13:20:36 - Il-2
Frames: 9921 - Time: 120000ms - Avg: 82.675 - Min: 43 - Max: 91 

LB

 

 

 

Ace !! 😀👍

Edited by SCG_Fenris_Wolf

Share this post


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

Second run on the GTX 1070 Max-Q

Frames Avg 30.225, Min 33, Max 44

 

Many thanks blitze for running the test. It is the first laptop in our table!!  Just curiosity, is there any reason for you to use a laptop for IL-2 VR?

 

It is also the first Pimax8K test. For the time being is OK to use 26% SS (like in Pimax5K+) but I need to calculate the exact number for the Pimax8K. Could you please give me two screenshoots of SteamVR Video SS: one with 26% and another with 100%?

 

Also, your test with the 1070 shows some strange numbers. If the Min is 33 the Avg can not be 30.225. Looking to the graph it seems you make a typo, should it be 40.225?

On 2/28/2019 at 11:59 PM, EAF51_Jimmi said:

Today i run the same test with HT disabled and CPU overcklocked to 4,9ghz - Test results were the same from the previous 4,3ghz test... any idea why i din't got any gain?

 

The cpu is capping the gpu, but increasing the oc didn't get results...

Tempertures were stable in the rage of 70°.... so i doubt ther's any throttling..

 

Throttling is not only for temp. It could be due to amps or power. If your temps are OK try to put higher limits of amps and power in BIOS.

 

You can also trend the CPU freq while you run the test using MSI afterbuner. Just to be sure there is not throttle.

 

You can also do multiple test form 4.0 to 4.9 with increments of 0.1. Just to see where you hit the max fps.

Share this post


Link to post
Share on other sites

Yeah, sorry about that.  The 1070 average was 39.225fps

 

Was a long day.

 

Images as requested the 1070 one just showing the SteamVR recommendation for the GPU.

 

Not much tweaking I can do with the i7700HQ - aside from undervolting.  I can overclock the GPU +200 on both but I run them stock and I hear memory overclocking on Video Cards now days doesn't give much benefit.

 

For the laptop, I spend time between countries with 3 and a half months in Aus.  Thought at the time I got it, the laptop would be ok for my duties.  It when hooked up to a monitor did decent speed on 1440p Il2.  It didn't do too bad for what it is in VR but at low settings for reasonable res and frames.  Your track is quite a cruncher - last night I was flying Stalingrad CAP missions and on the eGPU at 2160v res I was sitting between 50 to 70fps on my Balanced Setting.  Normal Antialiasing off but Nividia Antialiasing Transparency set to 4x SS which gives no frame hit but sharpens things up nicely.  I am still experimenting with settings though - I need some of that magic Fairy Dust.

 

The newer versions of my Laptop might also do better in CPU Mark as they are either a i9-8950HK or i7-8750H CPU's.

 

I also think, eGPU's will be better served with a more direct to CPU bus but there are security concers with that type of architecture - not that Intel doesn't have enough of those already including with Thunderbolt.

 

Anyway, not without it's issues but I am glad to have what I have both the laptop and Pimax 8K.  Not the rocking PC beast that some have but it puts up a reasonable effort.  I also still think that working with the native panel resolution of headsets or factors there of will give the best results in speed and image.  All the lens distortion geometry calcs and stuff are done before the info is fed into the GPU for rendering and then sending to the HMD.  I also think the scaler chip in the Pimax 8K might also hinder the image clarity and it would be great to eventually bypass that.

 

Anyway - VR is great - even with what we go through to deal with it.  Better than Monitor, better than Stereo Glasses - only to be surpassed by Holodecks - one day.

Test SteamVR App setting.jpg

Test SteamVR Video setting.jpg

Test 2 1070.jpg

Share this post


Link to post
Share on other sites

ok did more tests...

 

@4,6ghz   cpu mark 10584   single thread 2460
2019-03-02 23:53:51 - Il-2
Frames: 5198 - Time: 120000ms - Avg: 43.317 - Min: 30 - Max: 46

rtx 2070 1620mhz stock


@4,8ghz  cpu mark 10927   single thread 2510

a second run yelded better results cpu mark 10974 single thread 2582
                                                                                             

2019-03-03 00:16:38 - Il-2
Frames: 5337 - Time: 120000ms - Avg: 44.475 - Min: 35 - Max: 46
rtx 2070 1620mhz stock

2019-03-03 00:28:15 - Il-2
Frames: 5350 - Time: 120000ms - Avg: 44.583 - Min: 39 - Max: 46
rtx 2070 1800mhz  overcklock

 

i got gains in min FPS but non in maximum....

i was thinking the odissey was capped at 45.... but at high altitude i can reach 90fps in solo flying...

will check if motion reprojection is running...

Share this post


Link to post
Share on other sites

Checked the spreadsheet  all Odyssey are max capped at 46... 

i feel ther's something wrong in the test metodology with this headset...

Share this post


Link to post
Share on other sites
1 hour ago, EAF51_Jimmi said:

Checked the spreadsheet  all Odyssey are max capped at 46... 

i feel ther's something wrong in the test metodology with this headset...

 

Think the whole idea around this is not quite optimal. For measurements I still think we need FCAT VR and not “the last update in 2013” FRAPS. 

 

https://m.youtube.com/watch?v=VipgmnfrBMc

 

 

  • Upvote 1

Share this post


Link to post
Share on other sites

The only question is how do you compare data sets from different headsets with FCAT VR?

 

Do we agree upon a render target resolution with which to test a il2 track on? Something that has everyone rendering the same amount of pixels to HMD regardless of FOV

Share this post


Link to post
Share on other sites
11 hours ago, Dutch2 said:

Think the whole idea around this is not quite optimal. For measurements I still think we need FCAT VR and not “the last update in 2013” FRAPS. 

 

https://m.youtube.com/watch?v=VipgmnfrBMc

 

FCAT would certainly be better, but FRAFS does a decent enough job of showing a frame time plot. It helped me a lot in my VR tuning.

 

https://sourceforge.net/projects/frafsbenchview/

 

 

  • Like 1

Share this post


Link to post
Share on other sites
Posted (edited)

Just ran this benchmark.

 

System Specs

  • CPU: i5 9600K (listed as 3.7 GHZ but CPU-Z shows it running at 4.3 GHz, even at idle)
  • RAM: 16GB 2600 Mhz, 19-19-19-43 (I know, timings could be better)
  • GPU: GTX 1080
  • Storage: Samsung 970 EVO SSD 500 GB
  • HMD: Oculus Rift CV1
  • IL-2 Version: 3.010c

 

Benchmark Results

  • Frames: 7020
  • Avg: 58.500
  • Min: 42
  • Max: 88
  • CPU Mark score: 13549

 

Notes

  • When I set the SteamVR resolution to 150% the OTT overlay was reporting a pixel density of 1.23. To get the required 1.32 pixel density I had to use a SteamVR resolution of 172%. 172% was used for the benchmark.
  • Usually I fly with OpenComposite, pixel density 1.0, in-game AA 4x, ASW auto, and I'm happy with the performance.
  • I should really experiment with overclocking my CPU a bit. Everything is running at stock speeds for now. I have a Hyper T4 cooler, so there's probably some thermal headroom there.
  • I should have bought better RAM.
Edited by lightbulbjim
Added info.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...