Jump to content

SYN_Vander BENCHMARK v6 to measure IL-2 performance in monitor & VR


chiliwili69
 Share

Recommended Posts

10 minutes ago, chiliwili69 said:

 

Thanks for your test. Very much aligned with similar CPUs. Do you play in VR or just monitor?

 

I play online (TAW, WOL etc). Resolution 2560x1080.

Link to comment
Share on other sites

11 hours ago, Norz said:

Resolution 2560x1080

In that case you are quite OK with your PC. You will be above 60Hz most of the time if not all.

If one day you want to improve a bit, a faster RAM like 2666MHz will help.

Link to comment
Share on other sites

After a few weeks of reading and playing around with OC'ing my PC, I've now got a somewhat stable system (although still some issues with the DRAM), so here's my updated numbers for the CPU test,  and video test:

 

Motherboard: Gigabyte Aorus Z490 Elite
 CPU:                 i7-10700K
 CPU Freq:        4.8 Ghz
 L3 cache:        16 MB
 Cores:               8
 Threads:           16
 RAM type:        DDR4
 RAM size:        16Gb (2X8GB)
 NB Freq:          2400
 RAM Freq:        4400
 RAM timings:  19-19-19-39 770
 GPU:                 RX 5700 XT

 

Optional:

CPU Cooler: Fractal Design S24
RAM Model: Patriot Steel Viper 4400 Kit
GPU Model: Gigabyte OC

 

 

CPU test
2020-12-10 16:15:55 - Il-2
Frames: 5456 - Time: 60000ms - Avg: 90.933 - Min: 79 - Max: 125

 

GPU test

2020-12-10 16:25:12 - Il-2 @1440 monitor
Frames: 4834 - Time: 60000ms - Avg: 80.567 - Min: 70 - Max: 91

 

After various Memtest86 runs, the most stable memory was 3900Mhz.  I'm still contacting Patriot support as those Memtest runs revealed a possible bad stick of RAM.  We'll see if I can play with this a little later tonight without freeze ups in IL2.

Edited by Voxman
Link to comment
Share on other sites

7 hours ago, Voxman said:

NB Freq:          2400

Thank you Voxman for these new tests. I have seen your Uncore freq is 2400. Other people with an 10700K are using higher Uncore freqs in the order of 4000. Perhaps you can play with different profiles and pick the one with the highest NB freq and maybe lower RAM freq.

Link to comment
Share on other sites

6 hours ago, chiliwili69 said:

Thank you Voxman for these new tests. I have seen your Uncore freq is 2400. Other people with an 10700K are using higher Uncore freqs in the order of 4000. Perhaps you can play with different profiles and pick the one with the highest NB freq and maybe lower RAM freq.

Looks like more stuff to Google and learn.  Seriously, being away from this for eight years has been mind boggling.  So many ways and options to OC.  My previous CPU was an i5-2500K  with a very limited motherboard bios setup.  You just upped the frequency and voltage...I don't even think I had to overclock the RAM, or the settings were different on the mobo.  And the GPU is next...ha!

Link to comment
Share on other sites

Chili, curiosity got the best of me with your post (thank you, btw), and I Googled a little bit and played with my BIOS.  I changed the Ring Ratio to get close to the CPU ratio and now I'm getting a Uncore Fq of 4600.  I also disabled Hyperthreading as one thing I noticed with doing that was the reduction of thermal temps at higher OC clocks.  I did that and upped my OC to 4.9Ghz and ran Prime95 for over an hour or so.  Temps looked good and computer didn't lock up.  Then I had a little bit of time to play in game and ran HW monitor to check temps, and all that looked good.  There's a possibility I can OC this CPU a little bit more.  But that's for another day...or if curiosity gets the best of me. Many thanks on the suggestions!

  • Like 1
  • Thanks 1
Link to comment
Share on other sites

So my 6800 arrived.

 

Motherboard: Asus Crosshiar VIII Hero X570 (Bios 3303)

CPU: AMD Ryzen R9 5950X

CPU Frequency: 3.6Ghz

Level 3 Cache: 64MB

Cores: 16

Threads: 32

Ram Type: DDR4

Ram Size: 4 x 8GB

Uncore Frequency: 1866

Ram Frequency: 2 x 3733MHz

Ram Latency: 15 15 15 30 CR2

GPU: AMD Radeon RX6800 (Boost clock 2450Mhz - Ram 2100Mhz)

 

1440 GPU Test nearly maxed out of curiosity 

20201212_002140-L.jpg

 

20201212_002217-L.jpg

 

20201212_002242-L.jpg

 

20201212_002307-L.jpg

 

1440 Curiosity Test

2020-12-12 00:20:14 - Il-2
Frames: 4862 - Time: 60000ms - Avg: 81.033 - Min: 72 - Max: 92

 

VR Test 

2020-12-13 21:41:19 - Il-2
Frames: 5185 - Time: 60000ms - Avg: 86.417 - Min: 45 - Max: 91

 

 

Il-2 2020-12-12 00-20-14-44 fps.zip

 

Il-2 2020-12-13 21-41-19-86 fps.zip

Edited by robbiec
Link to comment
Share on other sites

14.11.2020 в 23:42, chiliwili69 сказал:

Pimax8KX  SS=158%   3158x3092=19.53 Mpixels (75HZ, with Normal FOV, Render Quality=1)

I want to do tests with 8kh, but 158% SS I see the resolution in StemVR much higher (parallel projections off.)

Link to comment
Share on other sites

WheelwrightPL

My VR TEST #2 (more intensive one) Results

 

HARDWARE:
- Motherboard: MSI B550-A PRO on latest BETA BIOS, XMP profile selected
- CPU: AMD Ryzen 7 5800X running on bigass 2-fan Noctua cooler
- RAM: XPG SPECTRIX D50 RGB 32GB (2x16GB) 3600MHz, timings: 3:54, 18, 20, 20, 42, 83
- GPU: ASUS ROG Strix GeForce RTX 2080TI (Factory) Overclocked 11G
- VR: HP Reverb G1 supersampled up to 206%, running on latest STEAMVR 1.15.12 with no reprojection

 

NVIDIA:
- Global Power Management Mode: Prefer Maximum Performance (there is no local IL2.exe Program Setting)

 

WINDOWS 10 OS:
- Windows OS Build# 19042.685 (latest)
- Update Paused
- Game Mode ON

 

IL2:
-FLIGHT INTERFACE: NORMAL
-REALISM: NORMAL
-AFTER LOADING MISSION: map hidden, icons hidden, waited 25secs before unpausing/recording


FRAPS RESULTS:

2020-12-13 12:13:54 - Il-2
Frames: 4109 - Time: 60000ms - Avg: 68.483 - Min: 57 - Max: 81

2020-12-13 12:53:52 - Il-2
Frames: 4082 - Time: 60000ms - Avg: 68.033 - Min: 55 - Max: 83

2020-12-13 12:57:09 - Il-2
Frames: 4086 - Time: 60000ms - Avg: 68.100 - Min: 52 - Max: 83

Link to comment
Share on other sites

WheelwrightPL

I noticed that SYN_VADER VR benchmarks use unrealistic (at least for me) graphics settings because almost everything is set to off.

Why would anybody buy $1500 graphics card and then a $1000 VR headest and then cripple his/her graphics like that ?

So I decided to run SYN_VADER VR benchmark using settings which I think more closely resemble what actual VR users use.

 

Therefore I did the following changes:

1) Upgraded graphics settings for optimal balance between VR performance and quality (at least for me), see the screenshot below.

2) Changed my HP Reverb G1 supersampling from 206% to 100% (because I see no need for this performance-sapping antialiasing).

 

With those  changes here are the (disappointing) FRAPS results I obtained:

2020-12-13 13:18:39 - Il-2
Frames: 2708 - Time: 60000ms - Avg: 45.133 - Min: 42 - Max: 49

2020-12-13 13:21:48 - Il-2
Frames: 2724 - Time: 60000ms - Avg: 45.400 - Min: 43 - Max: 51

2020-12-13 13:25:23 - Il-2
Frames: 2741 - Time: 60000ms - Avg: 45.683 - Min: 43 - Max: 51

 

 

 

 

MDGSFV.jpg

Link to comment
Share on other sites

WheelwrightPL

For reference I just re-ran the test on a monitor in 4k resolution at much higher graphics settings, here are my results:

 

2020-12-13 15:43:49 - Il-2
Frames: 3600 - Time: 60000ms - Avg: 60.000 - Min: 58 - Max: 62 (and BTW my monitor doesn't go higher than 60Hz, so I am effectively maxing-out)

 

Speaks for itself.

Edited by WheelwrightPL
Link to comment
Share on other sites

On 12/12/2020 at 1:35 AM, robbiec said:

VR Test 

2020-12-13 21:41:19 - Il-2
Frames: 5185 - Time: 60000ms - Avg: 86.417 - Min: 45 - Max: 91

Thank you for this. It is the first 6800 tested here. 

What VR test is it? 1 or 2?

On 12/12/2020 at 7:23 PM, WallterScott said:

I want to do tests with 8kh, but 158% SS I see the resolution in StemVR much higher (parallel projections off.)

Oh!  I have not an 8KX, so my numbers might be wrong then.

Could you tell me what resolution shows SteamVR for 100% in Normal FOV and RQ=1.

For the VR tests, look at the final number of pixels wo be around 9.5 Mpixels and 19.5.

4 hours ago, WheelwrightPL said:

Frames: 4086 - Time: 60000ms - Avg: 68.100 - Min: 52 - Max: 83

Thank you for your results. I would say you are gpu limited in these tests. (this was the purpose of this VRtest2)

 

3 hours ago, WheelwrightPL said:

noticed that SYN_VADER VR benchmarks use unrealistic (at least for me) graphics settings because almost everything is set to off

 

The settings of the tests are a bit strange, but they are like that because they are focused to stress just the CPU or the GPU, but no both.

 

The CPU test uses settings which stress the CPU (like shadows) and other settings like clouds are et to low because they stress the GPU. So we can know independenly which are the best CPU/RAMs for IL-2, regardeless of the GPU.

 

The GPU test is the same but loading GPU and not CPU.

 

For the VR test1, we want a value between 45 and 90, but no very close to 45 or very close to 90. So, we had to chose those settings to don´t stress the CPU too much.

The SYN_Vander flight is also not truly realistic, since 90% of your flying time you are not crossin a dense train station like that.

Link to comment
Share on other sites

4 hours ago, WheelwrightPL said:

ith those  changes here are the (disappointing) FRAPS results I obtained:

2020-12-13 13:18:39 - Il-2
Frames: 2708 - Time: 60000ms - Avg: 45.133 - Min: 42 - Max: 49

2020-12-13 13:21:48 - Il-2
Frames: 2724 - Time: 60000ms - Avg: 45.400 - Min: 43 - Max: 51

2020-12-13 13:25:23 - Il-2
Frames: 2741 - Time: 60000ms - Avg: 45.683 - Min: 43 - Max: 51

 

Yeap!, as I said. Those settings will not provide a useful information for a benchmark since they are lock to 45. Even with "only" 9.5 Mpixels and a 2080Ti. But here the bottleneck is still the CPU. So, new ryzens are good but VR can not deliver 90fps always at good settings and complex scenarios.

2 hours ago, WheelwrightPL said:

For reference I just re-ran the test on a monitor in 4k resolution at much higher graphics settings, here are my results:

 

2020-12-13 15:43:49 - Il-2
Frames: 3600 - Time: 60000ms - Avg: 60.000 - Min: 58 - Max: 62 (and BTW my monitor doesn't go higher than 60Hz, so I am effectively maxing-out

This effect was also commented by another person. My 4K monitor run at 60Hz but fraps can report fos higher than 60. Don´t know why.

Link to comment
Share on other sites

WheelwrightPL
1 hour ago, chiliwili69 said:

 

Yeap!, as I said. Those settings (which produce Avg: 45.133 fps) will not provide a useful information for a benchmark since they are lock to 45. Even with "only" 9.5 Mpixels and a 2080Ti. But here the bottleneck is still the CPU. So, new ryzens are good but VR can not deliver 90fps always at good settings and complex scenarios.

 

If my $700+ CPU is the bottleneck then I guess I will have to wait 10 years for 10Ghz+ CPUs, so I can jump the chasm between 45fps and 90fps (the latter is a must for smooth VR experience without reprojection-introduced graphical errors).

Link to comment
Share on other sites

Ii wonder is there a way we can utilise the SysInternal toolset and identify the cpu spikes? Perhaps the dev team could have a look? We are reporting that cpu type a can do an average of value a with this sort of map load, add gpu a to cpu a and it can scale resolution / feature wise and return value b, that pretends to be device HP Reverb G2 or anywhere in between.

 

@WheelwrightPL I think there is enough compute power cpu wise currently but the Il2 engine as it is currently, is not harnessing it fully. On the CPU bound test I'm maxing out at 5%, if I set my cpu to 8 Cores like your 5800X I'm seeing 11%, so IL2 is only using 3 or 4 Cores max. 

 

Link to comment
Share on other sites

8 hours ago, robbiec said:

2

 

I see you use a Rift CV1. (BTW, what a nice ergonomics had the CV1!!)

But 1880x2240x2=8.4 Mpixels, and for the VRtest1 the procedure was to do it with 9.5 Mpixels.

So I believe it would be 222% SS in SteamVR (or 1.49 PD in Oculus Tray Tool)

8 hours ago, WheelwrightPL said:

I guess I will have to wait 10 years for 10Ghz+ CPUs,

 

If in the last 6 years we went from 4.8GHz to 5.4GHz, I think you will need to wait more for the 10GHz.  😉

 

But this benchmark is just a benchmark. Something artficial to test individual performance of CPU and GPU.

You have one of the best CPUs to run IL-2 (you could run the CPU test just to verify everything is OK, and also VRtest1), it has been a good step from AMD and we are lucky with that. But, the IL-2 scenarios can be as dense as you wish. The SYN_vander benchmark is a well dense low alttitude scenario, specially in buildings. But a typical airfield is much less dense than that, and in you regular use of IL-2 you are 1000+ meters alttitude and you will be at 90fps most of the time.  

 

We have seen that Zen3 are a little bit better CPUs in monitor. We could extrapolate that to VR, but then the GPU enter in the scene. And then we have four combinations:

AMD CPU + AMD GPU

AMD CPU + NVIDIA GPU

Intel CPU + AMD GPU

Intel CPU + NVIDA GPU

 

So far in the VRTest1, we have 3 people above 85fps (Fenris with R95900X+3080, thermoregulator R95900X+3090, Hellfirelx R55600X+1080Ti). All of them are AMD CPU + NVIDIA GPU.

I want to see AMD GPUs also above 85fps in the VRTest1.

On 12/12/2020 at 1:35 AM, robbiec said:

CPU: AMD Ryzen R9 5950X

 

BTW, you could run again the CPU test because in your previous it could be limited by your old Vega64.

  • Thanks 1
Link to comment
Share on other sites

 

Well guys, I have also upgraded from a 4790K at 4.8 to a R5 5600X. All the upgrade Mobo+CPU+RAM+SSD+Win10 installation went well!. Keeping the 1080Ti, case, PSU and AIO 240 cooler.

 

I have just run the CPU test and GPU Test with stock settings (only activated the XMP profile of the RAM):

 

CPU Test with 5600X:

Frames: 7042 - Time: 60000ms - Avg: 117.367 - Min: 101 - Max: 158
Frames: 7046 - Time: 60000ms - Avg: 117.433 - Min: 101 - Max: 159
Frames: 7215 - Time: 60000ms - Avg: 120.250 - Min: 106 - Max: 167

 

As expected!

GPU Test with 5600X:
Frames: 5677 - Time: 60000ms - Avg: 94.617 - Min: 75 - Max: 110
Frames: 5714 - Time: 60000ms - Avg: 95.233 - Min: 78 - Max: 108

 

As expected!

 

My old results with the overclocked 4790K in the CPU test (80.3 fps) and GPU test (93.62 fps). So nice boost on the CPU side.

 

My new specs:

 

 Motherboard: Gigabyte B550 Aorus Pro
 CPU:                 R5 5600X
 CPU Freq:        4.65 Ghz  (Not fixed to a value, in Auto it went from 3.1 to 4.65)
 L3 cache:        32 MB
 Cores:               6 
 Threads:           12 
 RAM type:        DDR4
 RAM size:        32Gb -  G.Skill Trident Z 2x16GB 3600 (F4-3600C16D-32GTZR)
 NB Freq:          1800 MHz (or Uncore Frequency or Infinity Fabric freq)
 RAM Freq:        3600 MHz (this is 2xDRAM freq if Dual channel)
 RAM timings:  16-16-16-36-85
 GPU:                 1080Ti

Edited by chiliwili69
  • Like 2
  • Upvote 1
Link to comment
Share on other sites

6 minutes ago, chiliwili69 said:

 

Well guys, I have also upgraded from a 4790K at 4.8 to a R5 5600X. All the upgrade Mobo+CPU+RAM+SSD+Win10 installation went well!. Keeping the 1080Ti, case, PSU and AIO 240 cooler.

 

 

Looking good. Waiting for the 3080 Ti..?

 

Link to comment
Share on other sites

7 hours ago, chiliwili69 said:

 

My old results with the overclocked 4790K in the CPU test (80.3 fps) and GPU test (93.62 fps). So nice boost on the CPU side.

Very nice bump!

Link to comment
Share on other sites

8 hours ago, chiliwili69 said:

So far in the VRTest1, we have 3 people above 85fps (Fenris with R95900X+3080, thermoregulator R95900X+3090, Hellfirelx R55600X+1080Ti). All of them are AMD CPU + NVIDIA GPU.

I want to see AMD GPUs also above 85fps in the VRTest1.

 

So given that, for those of us who have 1080 Ti's already, does it actually even make sense to upgrade to this gen for VR? 

 

Or are we best off just keep the 1080 TI and focusing on getting the CPU upgraded? 

Link to comment
Share on other sites

1PL-Husar-1Esk
16 minutes ago, Voyager said:

 

So given that, for those of us who have 1080 Ti's already, does it actually even make sense to upgrade to this gen for VR? 

 

Or are we best off just keep the 1080 TI and focusing on getting the CPU upgraded? 

They have high end GPU plus modern CPU,  I doubt there will be significant difference between ryzen 5900x and i9 10900k with those GPUs. 1080ti is not enough for 90 fps vr in HP G2. You need both good modern CPU and better GPU. 

Edited by 1PL-Husar-1Esk
Link to comment
Share on other sites

25 minutes ago, 1PL-Husar-1Esk said:

They have high end GPU plus modern CPU,  I doubt there will be significant difference between ryzen 5900x and i9 10900k with those GPUs. 1080ti is not enough for 90 fps vr in HP G2. You need both good modern CPU and better GPU. 

 

That's a different question than what I was asking. 

 

Basically, if the 1080 Ti already gets 85+ fps in VR Test 1, when paired with a Zen 3 chip, how big an improvement does a 3080 or 3090 actually bring to the table? 

 

I've been going back and forth on what CPU/GPU to replace my 3800X/1080 Ti with, but if a Zen 3/1080 TI gets me 60fps+ performance as is, I'm really wondering if I should just grab a 5800X when they come back in stock and just sit this GPU generation out? 

Link to comment
Share on other sites

Hardware

 Motherboard:    Gigabyte X570 Aorus Pro Wifi
 CPU:        AMD Ryzen 7 3800X
 CPU Freq:    4450 MHz
 L3 cache:    2x16 MB
 Cores:        8
 Threads:    16
 RAM type:    DDR4
 RAM size:    64 GB (4x16)
 NB Freq:    1796.4 MHz
 RAM Freq:    1796.4
 RAM timings:    16-19-19-39-58
 GPU:        GTX 1080 Ti

IL:-2 version 4.503

 

VR Headset: HP Reverb G1

 

Optional:

CPU Cooler: Corsair Hydro H150
RAM Model: G.Skill SK Hynix F4-3600C16-16GVKC (Dual Rank)
GPU Model: Gigabyte GTX 1080 Ti FE

 

CPU Test
2020-12-15 00:47:13 - Il-2
Frames: 5125 - Time: 60000ms - Avg: 85.417 - Min: 73 - Max: 118

 

Pseudo GPU Test (DQHD, 5120x1440p: 7.37m pixels)
2020-12-15 00:56:04 - Il-2
Frames: 5374 - Time: 60000ms - Avg: 89.567 - Min: 72 - Max: 103

 

VR Test 1

HP Reverb G1 SS: 100%

2020-12-15 01:01:29 - Il-2
Frames: 3282 - Time: 60000ms - Avg: 54.700 - Min: 45 - Max: 74

 

VR Test 2

HP Reverb G1 SS: 206%

2020-12-15 01:07:17 - Il-2
Frames: 2899 - Time: 60000ms - Avg: 48.317 - Min: 42 - Max: 57

 

I don't have a 4k monitor, but just to see what happened, I ran it on a DQHD 5120x1440p display to see how it did. 

Edited by Voyager
Added missing hardware info
Link to comment
Share on other sites

23 hours ago, Pinello said:

Waiting for the 3080 Ti..?

 

Don´t know really. I would like to see more results for 3070, 6800XT and 3080. And more reasonable prices.

16 hours ago, Voyager said:

does it actually even make sense to upgrade to this gen for VR? 

 

Or are we best off just keep the 1080 TI and focusing on getting the CPU upgraded?

 

When you say "this gen of VR" I suposse you refer to G2 or Pimax8KX.

 

For me hte CPU upgrade was the first priority, since it was the main constraint for the Index right now. The extra fps given by the new Zen3 outofthebox are quite good.

Then the second question is the GPU. On that area there is a lot of area to experiment and analyze before having a clear choice.

Link to comment
Share on other sites

2 hours ago, Voyager said:

VR Test 1
2020-12-15 01:01:29 - Il-2
Frames: 3282 - Time: 60000ms - Avg: 54.700 - Min: 45 - Max: 74

 

VR Test 2
2020-12-15 01:07:17 - Il-2
Frames: 2899 - Time: 60000ms - Avg: 48.317 - Min: 42 - Max: 57

 

Thanks for this. One the VRTest1 the CPU is the bottleneck. On the VRTest2 the CPU+GPU is the bottleneck.

BTW, which is your VR device and SS applied in each test?

 

Let´s see if I can find time to run the two VR tests just with stock settings.

Link to comment
Share on other sites

@chiliwili69 Sorry about that. I've added the headset and SS to the list: HP Reverb G1 at 100% then 206% per test instructions. 

 

On GPU for VR, I was intending to be asking if a 30 series or a RDNA2 would be a meaningful improvement for HP Reverb G1 class devices, when paired with a Zen 3 CPU.

 

My thing is I've got an oversized head and 71mm IPD so I'm probably going to pass on the G2, and might upgrade if the Decagear one actually pans out, but that's going to be a Reverb G1 scale device, not a G2 one. 

 

If the 1080 TI can get me consistent good fps in the Reverb G1 it may make more sense for me to hold onto it for another year or two. 

Link to comment
Share on other sites

23 hours ago, Voyager said:

That's a different question than what I was asking. 

 

Basically, if the 1080 Ti already gets 85+ fps in VR Test 1, when paired with a Zen 3 chip, how big an improvement does a 3080 or 3090 actually bring to the table? 

 

I've been going back and forth on what CPU/GPU to replace my 3800X/1080 Ti with, but if a Zen 3/1080 TI gets me 60fps+ performance as is, I'm really wondering if I should just grab a 5800X when they come back in stock and just sit this GPU generation out? 

 

I had a 2080 (=1080ti) and I upgraded to a 3080. In my real-world usage, which is focused on multiplayer, I was able to go from shadows off to shadows ultra, grass off to grass ultra, and clouds medium to clouds high (all on High preset, no change there). I also went from FXAAx4 to MSAAx4 and SteamVR supersample from 100% to 110%. All while maintaining a pretty solid 80 FPS. Drops below 80 are mostly due to my CPU, which is "only" a 5.1ghz 8086K.

 

If you are running your headset at 60hz and are happy with that, and your CPU currently delivers enough power for the 60 FPS, then you don't need to change it. If you buy a better headset and need to render more pixels, or if you want to increase GPU-based eye candy (shadows, clouds) then you might consider a GPU upgrade. But it's not going to be some kind of night-and-day difference. Personally I really like being able to enable shadows again and the shift from FXAA to MSAA has been really nice, but it's still the same game really.

  • Upvote 1
Link to comment
Share on other sites

Here my VR test 1 and 2 with the Index at 90Hz

 

VRtest1: SS106% : 9.5 million pixels

Frames: 4574 - Time: 60000ms - Avg: 76.233 - Min: 44 - Max: 91
Frames: 4653 - Time: 60000ms - Avg: 77.550 - Min: 46 - Max: 91
Frames: 4749 - Time: 60000ms - Avg: 79.150 - Min: 48 - Max: 91

 

Here a nice boost over my previous 45fps with the old rig. So CPU upgrade helps.

The fluctuations on the fps might be due to Auto stock settings (CPU freq varies from 3.3 to 4.65).

 

VRtest2: SS216% : 19.5 Million pixels

Frames: 2720 - Time: 60000ms - Avg: 45.333 - Min: 42 - Max: 59
Frames: 2729 - Time: 60000ms - Avg: 45.483 - Min: 42 - Max: 59
Frames: 2749 - Time: 60000ms - Avg: 45.817 - Min: 42 - Max: 60

 

Here clearly the 1080Ti can not handle the equivalent pixels of the G2 at 100%

 

Now it is time to experiment a little with overclocking CPU and a bit RAM and IFCLOCK freq.  This is completely different from overclocking intel CPUs (it was so simple).

I installed Ryzen Master which looks pretty nice, but Fenris recommended to not use it and use the BIOS.

 

 

 

 

Link to comment
Share on other sites

Would it be reasonable to use this benchmark to dial in settings for specific minimum frame rates? 

 

I'm wanting to run with the view distanced maxed, for high altitude pilotage, and I'm wondering if I could use this to dial things in to get a consistent minimum 45 fps? 

Link to comment
Share on other sites

1 hour ago, chiliwili69 said:

Here my VR test 1 and 2 with the Index at 90Hz

 

VRtest1: SS106% : 9.5 million pixels

Frames: 4574 - Time: 60000ms - Avg: 76.233 - Min: 44 - Max: 91
Frames: 4653 - Time: 60000ms - Avg: 77.550 - Min: 46 - Max: 91
Frames: 4749 - Time: 60000ms - Avg: 79.150 - Min: 48 - Max: 91

 

Here a nice boost over my previous 45fps with the old rig. So CPU upgrade helps.

The fluctuations on the fps might be due to Auto stock settings (CPU freq varies from 3.3 to 4.65).

 

VRtest2: SS216% : 19.5 Million pixels

Frames: 2720 - Time: 60000ms - Avg: 45.333 - Min: 42 - Max: 59
Frames: 2729 - Time: 60000ms - Avg: 45.483 - Min: 42 - Max: 59
Frames: 2749 - Time: 60000ms - Avg: 45.817 - Min: 42 - Max: 60

 

Here clearly the 1080Ti can not handle the equivalent pixels of the G2 at 100%

 

Now it is time to experiment a little with overclocking CPU and a bit RAM and IFCLOCK freq.  This is completely different from overclocking intel CPUs (it was so simple).

I installed Ryzen Master which looks pretty nice, but Fenris recommended to not use it and use the BIOS.

 

 

 

 

 

I would highly recommend the same for overclocking CPU on any board.

Only do it through the bios.

Link to comment
Share on other sites

13 hours ago, chiliwili69 said:

 

Don´t know really. I would like to see more results for 3070, 6800XT and 3080. And more reasonable prices.

 

 

 

Some very interesting benchmarks 3080 vs. 6800xt with x5900/x5600 in VR over at "the other sim". Threads by JayRoc, guess you have seen them..

Edited by Pinello
.
Link to comment
Share on other sites

Thanks Chili. Currently running a 4790k and a 2080 so AMD's new chips seem to make a big difference. Problem is they can't be got atm. Same for 6800xt. 

Link to comment
Share on other sites

Decided to give my 1440p monitor a go with the GPU test. Granted, it is not the requested 4K resolution, but since others are posting their 2K experience, I thought I would contribute. Also, this is with a RTX 3070, so some additional insight into this particular card's performance. I have included my last CPU test as well.

 

Motherboard: Gigabyte X570 Aorus Elite
CPU:                 AMD Ryzen 5 5600X
CPU Freq:        4.6 GHz
L3 cache:        32 MB
Cores:              6
Threads:          12
RAM type:       DDR4
RAM size:        32 GB (2 x 16GB)
RAM Freq:       3600 MHz 
RAM Latency: 16-18-18-38
GPU:                 Gigabyte Gaming OC RTX 3070

 

CPU TEST

2020-12-02 16:58:20 - Il-2
Frames: 7083 - Time: 60000ms - Avg: 118.059 - Min: 99 - Max: 188

 

GPU TEST (1440p)

2020-12-15 17:19:17 - Il-2
Frames: 8554 - Time: 60000ms - Avg: 142.567 - Min: 119 - Max: 146

 

Link to comment
Share on other sites

41Sqn_Skipper
On 12/13/2020 at 7:05 PM, WheelwrightPL said:

My VR TEST #2 (more intensive one) Results

...
- GPU: ASUS ROG Strix GeForce RTX 2080TI (Factory) Overclocked 11G
- VR: HP Reverb G1 supersampled up to 206%, running on latest STEAMVR 1.15.12 with no reprojection

...

FRAPS RESULTS:

2020-12-13 12:13:54 - Il-2
Frames: 4109 - Time: 60000ms - Avg: 68.483 - Min: 57 - Max: 81

2020-12-13 12:53:52 - Il-2
Frames: 4082 - Time: 60000ms - Avg: 68.033 - Min: 55 - Max: 83

2020-12-13 12:57:09 - Il-2
Frames: 4086 - Time: 60000ms - Avg: 68.100 - Min: 52 - Max: 83

 

Thanks for the test! The results are brutal. Even with most graphic features off or low it's not even close to 90fps. G2@100% resolution and 90Hz is basically impossible except for 3080/3090 (and even then requires to disable most graphic features). 

Link to comment
Share on other sites

1 hour ago, 41Sqn_Skipper said:

 

Thanks for the test! The results are brutal. Even with most graphic features off or low it's not even close to 90fps. G2@100% resolution and 90Hz is basically impossible except for 3080/3090 (and even then requires to disable most graphic features). 

 

Yep. I have a 3090 and run with Motion Smoothing enabled and 100% resolution and am quite happy.

Sim is gorgeous in VR. I have graphics just about maxed out except no AA. 

Edited by dburne
Link to comment
Share on other sites

WheelwrightPL
3 hours ago, 41Sqn_Skipper said:

 

Thanks for the test! The results are brutal. Even with most graphic features off or low it's not even close to 90fps. G2@100% resolution and 90Hz is basically impossible except for 3080/3090 (and even then requires to disable most graphic features). 

 

It gets worse: look at my VR test right below this one which has more realistic settings (because the graphics are at medium/high as opposed to most-off, and the Reverb G1 supersampling has been dialed back to 100% from unrealistic 206%).

That second VR test has the framerates in the 45fps range, or even lower, and  I am told I am CPU-limited by "only" $700 Ryzen 7 5800x. There is no hope for VR in its current state if you want relatively distortion-free (no reprojection) and smooth image with medium/high graphics on a decent-resolution headset.

  • Like 1
Link to comment
Share on other sites

On 11/20/2020 at 10:56 PM, Hartigan said:

Motherboard: Asrock Z270 Pro4
CPU:               I5 7600K
CPU Freq:      5.0 Ghz
L3 cache:       8 MB
Cores:             4 
Threads:         4
RAM type:       DDR4
RAM size:       16 GB
NB Freq:          3800 MHz
RAM Freq:       3000 MHz 
RAM Latency: 16
GPU:                  1080Ti 11Gb

 

 

CPU 5GHZ 
2020-11-17 17:29:04 - Il-2
Frames: 5173 - Time: 60000ms - Avg: 86.217 - Min: 76 - Max: 115

 

 

 

I have made new CPU test using my new RAM sticks ...waiting new mobo to arrive . 

Old sticks Ripjaws Aegis 3000mHz xmpProfile/ 2x8GB / timings 16/18/18/38 AND CPU 5GHZ . You can see test results above.

New Ripjaws is 3600mHz xmpPriofile /2x8GB/timings 16/16/16/36 AND CPU 4.8GHZ

Results :

 2020-12-16 17:10:39 - Il-2
Frames: 5393 - Time: 60000ms - Avg: 89.883 - Min: 79 - Max: 126

Note CPU from 5 to 4.8

Edited by Hartigan
Link to comment
Share on other sites

6 hours ago, WheelwrightPL said:

That second VR test has the framerates in the 45fps range, or even lower, and  I am told I am CPU-limited by "only" $700 Ryzen 7 5800x. There is no hope for VR in its current state if you want relatively distortion-free (no reprojection) and smooth image with medium/high graphics on a decent-resolution headset.

 

That's 45 FPS on the VR torture test, though. Fenris is running a 5900X (or maybe 5950 I'm not sure) and getting a locked 90 FPS for his use case, which is competitive multiplayer. Everything depends on the specific use case. If you want to run a single player mission with a ton of Ai that was created by folks who literally say they discount VR when designing missions (I'm not knocking them, they make great stuff it just has a lot of Ai) then you're probably going to be frame limited. But if you run a more conservative mission you might be fine.

 

But it's horses for courses. In IL2 multiplayer I'm hoping to be able to run a G2 at reasonable resolution and 90 FPS, and I know I might need a CPU upgrade to do it. In the other sim, which is much worse on the CPU but has a lot of BVR engagements, I'll be aiming for 45 with reprojection, and in fact I already know I can't run the G2 at anything like native resolution with "only" 10GB in my 3080. Will need a 20GB 3080ti to get that closer to 100%. But there it's a VRAM and CPU limitation.

Link to comment
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
 Share

×
×
  • Create New...