Jump to content

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


chiliwili69
 Share

Recommended Posts

52 minutes ago, Alonzo said:

 

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.

 

Just as an FYI to anyone, following the advice at https://www.reddit.com/r/nvidia/comments/iyl91r/comment/g6dic16?utm_source=share&utm_medium=web2x&context=3 , on Combat Box just now with about 40 other pilots the MSI Afterburner 4.6.3 Beta 2 Build 15840 shows I maxed out at 12.5 Gbs of VRAM "dedicated memory usage" (which I interpret as "allocated" and matches what fpsVR displays for me), but I maxed out at 7.1Gbs of VRAM "dedicated memory usage\process", which I interpret as actual VRAM used by IL2. This includes during shooting down an A20 low over Bochum with perhaps three other aircraft I could see around me.

  • Like 1
Link to comment
Share on other sites

On 12/15/2020 at 11:17 PM, Pinello said:

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

 

I dind´t know. IL-2 VR is my only sim or game. (well I will give a try on Alyx this weekend now I have controllers from G2).

He say that 6800XT is better than 3080 using AMD CPUs. We will need to see how it behaves in IL-2 VR.

23 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.

Hey it is not as tragic as you think. This benchmark is really really CPU intensive. It is not what you normally do in a normall mission. It is representing the 1% heaviest moment of your total flying time. With your G1 at 100%SS, 5800X and 2080Ti you will be at 90fps most of the time with your usual settings.

Perhaps CPU OC or RAM clock (IFCLK) can bring a few more fps.

23 hours ago, Hartigan said:

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

 

Thanks for this. What version of IL-2 did you used 4.503 or the last 4.504?

  • Upvote 1
Link to comment
Share on other sites

WheelwrightPL
20 minutes ago, chiliwili69 said:

Hey it is not as tragic as you think. This benchmark is really really CPU intensive. It is not what you normally do in a normall mission. It is representing the 1% heaviest moment of your total flying time. With your G1 at 100%SS, 5800X and 2080Ti you will be at 90fps most of the time with your usual settings.

Perhaps CPU OC or RAM clock (IFCLK) can bring a few more fps.

 

I am afraid that's incorrect. I tested a simple ground-attack Quick Mission in HS-129 on winter Velikie Luki map, and even during takeoff when I looked sideways at the forest it was stuttering very badly all the time. I had reprojection turned off. So I saw no visible improvement from upgrading to Ryzen 5800X and faster RAM & architecture. I suppose the reason is because jumping this huge chasm between 45fps to 90fps is simply not achievable with my graphics settings.

Link to comment
Share on other sites

40 minutes ago, chiliwili69 said:

 

I dind´t know. IL-2 VR is my only sim or game. (well I will give a try on Alyx this weekend now I have controllers from G2).

 

 

Not to get off topic, just wanted to comment I did a complete second play through of Alyx with the Reverb G2 and it was an incredible experience.

Absolutely loved it.

Link to comment
Share on other sites

Currently, there is a "ground roughness" installation. How do I set it up for CPU and virtual reality testing?

 

14.12.2020 в 01:19, chiliwili69 сказал:

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

 

RQ 1, PP off, normal fov, SS 80% (3324x2824) = 9,386Mp   SS 166% (4788x4068) = 19,463 (for Pimax 8kx)

 

 Motherboard: Asus Rog Maximus XII APEX
 CPU:                10900k
 CPU Freq:        5.4 Ghz
 L3 cache:        20 MB
 Cores:              10 (number of active physical Cores)
 Threads:          20 (HyperThreading is better off. If Cores=Threads then HT is Off, But as you wish)
 RAM type:       DDR4
 RAM size:       32 GB
 NB Freq:          4900 MHz (or Uncore Frequency)
 RAM Freq:       4200 MHz (this is 2xDRAM freq if Dual channel)
 RAM Latency: 16  (CAS Latency CL)
 GPU:                 3090 Strix OC 

 

VR 1 Test

2020-12-17 17:30:15 - Il-2
Frames: 4501 - Time: 60000ms - Avg: 75.017 - Min: 73 - Max: 77

 

VR 2 Test

2020-12-17 17:47:39 - Il-2
Frames: 4379 - Time: 60000ms - Avg: 72.983 - Min: 61 - Max: 76

 

Edited by WallterScott
Link to comment
Share on other sites

7 hours ago, chiliwili69 said:

He say that 6800XT is better than 3080 using AMD CPUs. We will need to see how it behaves in IL-2 VR.

 

I would take the benchmark numbers from that other sim with a pinch of salt. The person doing the benchmarks is very well meaning, but they confuse pixel density with SteamVR SS percentage and they're giving frame time ranges rather than anything that makes me think they did a Fraps recording. And the 6800XT is coming out ahead of the 3080 where for almost every other game at 4K or above, the 3080 is in the lead. Might be a VRAM thing since that other game is hideously badly optimized.

 

6 hours ago, WheelwrightPL said:

I am afraid that's incorrect. I tested a simple ground-attack Quick Mission in HS-129 on winter Velikie Luki map, and even during takeoff when I looked sideways at the forest it was stuttering very badly all the time. I had reprojection turned off. So I saw no visible improvement from upgrading to Ryzen 5800X and faster RAM & architecture. I suppose the reason is because jumping this huge chasm between 45fps to 90fps is simply not achievable with my graphics settings.

 

Not to beat a dead horse, but you need to get the computer to tell you why it's missing the 11ms frame time limit. Get FpsVR, run it during these problem areas, get it to tell you whether it's CPU or GPU time that is the problem. And, no, if you stick everything on Ultra and a G2 on 100% rendering, you're not going to get 90 FPS.

  • Upvote 2
Link to comment
Share on other sites

On 12/15/2020 at 1:11 AM, Voyager said:

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. 

 

 

Retest of system with 4.504

 

CPU Test
Terrain Roughness Off

2020-12-17 16:43:40 - Il-2
Frames: 5111 - Time: 60000ms - Avg: 85.183 - Min: 74 - Max: 117

 

Terrain Roughness High

2020-12-17 16:38:55 - Il-2
Frames: 5072 - Time: 60000ms - Avg: 84.533 - Min: 73 - Max: 117

 

Pseudo GPU Test (DQHD, 5120x1440p: 7.37m pixels)

Terrain Roughness Off (Only option)

2020-12-17 16:49:51 - Il-2
Frames: 5335 - Time: 60000ms - Avg: 88.917 - Min: 74 - Max: 102

 

VR Test 1

HP Reverb G1 SS: 100%

Terrain Roughness Off

2020-12-17 16:55:33 - Il-2
Frames: 3358 - Time: 60000ms - Avg: 55.967 - Min: 46 - Max: 80

 

Terrain Roughness High

2020-12-17 16:59:10 - Il-2
Frames: 3325 - Time: 60000ms - Avg: 55.417 - Min: 46 - Max: 77

 

VR Test 2

HP Reverb G1 SS: 206%

Terrain Roughness Off

2020-12-17 17:04:56 - Il-2
Frames: 2857 - Time: 60000ms - Avg: 47.617 - Min: 42 - Max: 57

 

Terrain Roughness High

2020-12-17 17:11:03 - Il-2
Frames: 2868 - Time: 60000ms - Avg: 47.800 - Min: 41 - Max: 54

 

 

Overall the test results for 4.504 look, to me, to be comparable with the 4.503 results, and setting the Terrain Roughness to High has minimum impacts on framerates for this test.

Link to comment
Share on other sites

17 hours ago, WheelwrightPL said:

simple ground-attack Quick Mission in HS-129 on winter Velikie Luki map, and even during takeoff when I looked sideways at the forest it was stuttering very badly all the time

 

I don´t fly the HS-129 (Don´t know how demanding is it) but I am intrigued with your performance with your typical settings. 

You have an 2080Ti, but I will try to run it at MSAAx2 (not x4) or better FXAAx2 (FXAA loads much less GPU).

 

Please, run the CPU test (and GPU test in 4K possible) of the benchmark just to be sure that your rig is running OK.

13 hours ago, WallterScott said:

SS 80% (3324x2824) = 9,386Mp   SS 166% (4788x4068) = 19,463 (for Pimax 8kx)

 

Thank walter for this information, but I don´t understand well those numbers.

When I say 9.5Million pixelsI referred to the total number of pixels of both eyes so it should be 2*H*V, being H the horizontal pixels and V the vertical pixels  of one eye shown by SteamVR.

Just my understanding, could you show a screenshoot of SteamVR manual SS setting with 100%SS?

9 hours ago, Voyager said:

Overall the test results for 4.504 look, to me, to be comparable with the 4.503 results, and setting the Terrain Roughness to High has minimum impacts on framerates for this test.

 

Thank you very much Voyager for doing this testing. It is something we will need to check for every update of the sim, just to compare the tests apple with apple.

 

I will introduce into the instructions to run the "Terrain roughness" OFF since it will be more consistent with previous test where this option was not available.

Link to comment
Share on other sites

40 минут назад, chiliwili69 сказал:

Just my understanding, could you show a screenshoot of SteamVR manual SS setting with 100%SS?

 

777.PNG

 

 

Then 19.5 megapixels turns out like this

 

888.PNG

Edited by WallterScott
Link to comment
Share on other sites

WheelwrightPL
7 hours ago, chiliwili69 said:

 

I don´t fly the HS-129 (Don´t know how demanding is it) but I am intrigued with your performance with your typical settings. 

You have an 2080Ti, but I will try to run it at MSAAx2 (not x4) or better FXAAx2 (FXAA loads much less GPU).

 

Please, run the CPU test (and GPU test in 4K possible) of the benchmark just to be sure that your rig is running OK.

 

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
 

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 OFF

 

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

 

 

CPU TEST (1080p) Results:

 

2020-12-18 10:17:26 - Il-2
Frames: 7526 - Time: 60000ms - Avg: 125.433 - Min: 112 - Max: 171

2020-12-18 10:20:16 - Il-2
Frames: 7443 - Time: 60000ms - Avg: 124.050 - Min: 111 - Max: 168

2020-12-18 10:32:59 - Il-2

Frames: 7391 - Time: 60000ms - Avg: 123.183 - Min: 107 - Max: 164

 

 

GPU TEST (4k) Results:

 

2020-12-18 11:01:25 - Il-2
Frames: 4301 - Time: 60000ms - Avg: 71.683 - Min: 59 - Max: 80

2020-12-18 11:04:15 - Il-2
Frames: 4288 - Time: 60000ms - Avg: 71.467 - Min: 59 - Max: 80

2020-12-18 11:10:09 - Il-2
Frames: 4301 - Time: 60000ms - Avg: 71.683 - Min: 59 - Max: 81

 

 

CPU TEST SETTINGS (Terrain roughness was also set to OFF):

CPU_test.jpg

 

 

GPU TEST SETTINGS (Terrain roughness was also set to OFF):

GPU_test.jpg

Edited by WheelwrightPL
Link to comment
Share on other sites

Hi,

 

I came back to this because I discovered that the GPU overclock was not stable and also because I made a big trial of the capacities of my systems in terms of PBO2 curve optimizer. Sadly I must admit that I couldn't really find  stable settings to my satisfaction. When it was stable in multiple core stress test where it ran 4550 all cores, it failed in single core, where it always hit 4850 on the 2 prefered cores. The problem is Windows distributes the single thread among all cores, and I couldn't get the offsets on individual cores to play along. Long story short, there is potential but I prefer to wait for the next BIOS upgrades. This is the same reason why i also disabled the 1900IF on DRAM.  I must also underline that I do base the instability status on OCCT CPU Stress Test Extreme with AVX Instructions. It usually produced errors in the 2 first minutes. Probably most applications or games wouldn't get into those limits anyway. Thanks to the 240mm AIO, the temps never really left the low 70's. I honestly think that it does not make much sense to try too hard, the Ryzen 5000 series is really generous in base settings.

 

This are the last benchmarks I will have with the 1080ti, because I couldn't resist and got my hands on a 3080 coming next week. @chiliwili69, could you please replace my actual results with the 5600x and 1080ti with these ?

 

Hardware (This settings are stable under stress for a prolonged time):

Motherboard:      MSI B550 Gaming Carbon Wifi
CPU:                     R5 5600X
CPU Freq:            max 4.85 Ghz - PBO advanced enabled,  PBOlimits motherboard, +200Mhz, no curve optimization.
L3 cache:             32 MB
Cores:                  6
Threads:             12
RAM type:          DDR4
RAM size:           32Gb (2x16GB)
NB Freq:             1800 MHz (or Uncore Frequency or Infinity Fabric freq), 1.050V
RAM Freq:         3600 MHz Dual Channel
RAM timings:    16-16-16-36 DRAM Stock
GPU:                    1080Ti - 1974 Mhz - Memory 5616 Mhz. Power limit to 150% , Temp limit to 86C.

 

Optional:

CPU Cooler: Arctic Liquid Freezer 240 AIO.
RAM Model: G-SKILL Trident Z Neo F4-3600C16D-32GTZN,  
GPU Model: Gigabyte AORUS GTX 1080 Ti Xtreme Edition 11G

 

CPU Test:

2020-12-18 16:58:52 - Il-2
Frames: 6862 - Time: 60000ms - Avg: 114.367 - Min: 100 - Max: 152

 

VR Test 1 - Index 106x

2020-12-18 21:42:42 - Il-2
Frames: 5241 - Time: 60000ms - Avg: 87.350 - Min: 79 - Max: 91

 

VR Test 2 - Index 216x

2020-12-18 21:48:07 - Il-2
Frames: 4266 - Time: 60000ms - Avg: 71.100 - Min: 58 - Max: 77

 

 

 

 

Link to comment
Share on other sites

7 hours ago, WallterScott said:

Then 19.5 megapixels turns out like this

 

Yes, you are right, to get 19.5 million you need to put 83% SS.

And to have 9.5 Million you need to run it at 40%. 

 

Pimax8KX.png.5d09cde5ed8c9f01f6687830fb47c92d.png

 

I will update the instructions then for the 8KX

  • Upvote 1
Link to comment
Share on other sites

SCG_Fenris_Wolf

Update for my (formerly incomplete) stats:

 

tRFC: 370

VR HMD: Reverb G2

CPU Cooler: Deepcool Castle 360 RGB V2

RAM: G.Skill Trident Z Neo 2x16GB 3600MHz CL16-16-16-36

 

 

I still didn't get to run the 4K test. It's good that you have a 19.5M p test now, I'll have to run that tomorrow.

Edited by SCG_Fenris_Wolf
Link to comment
Share on other sites

On 12/17/2020 at 8:18 PM, WallterScott said:

SS 80% (3324x2824) = 9,386Mp

 

This is the SS you applied for the VRTest1, but this is alsmot 19.5 million pixels (3324x2824x2), so I will take it as the VRTest2.

It is indeed a very good result for the 3090. You reached the 75fps limit because your Pimax 8KX runs at 75Hz.

 

In your second test you went quite beyond the 19.5 million pixels. you did it at SS 166% (4788x4068x2) which is 38.9 million pixels, and you got 73fps avg. That´s really amazing!!!

That 3090 is insane.

Link to comment
Share on other sites

Только что, chiliwili69 сказал:

so I will take it as the VRTest2.

I will do the tests again, with the correct values

Только что, chiliwili69 сказал:

That´s really amazing!!!

Still, when the fps dropped to 61-65, there were noticeable stutters. Probably 70 Hz is the minimum for smoothness. and in the test with 80% ss they were. Very short but I've seen stutters. I'll take a look with the fpsVR application running.

a 3090 is not a stock one. This is a Strix with a 520W bios, cooled by a waterchiller. here's the secret to performance))

Link to comment
Share on other sites

20 hours ago, WheelwrightPL said:

GPU TEST (4k) Results:

 

2020-12-18 11:01:25 - Il-2
Frames: 4301 - Time: 60000ms - Avg: 71.683 - Min: 59 - Max: 80

2020-12-18 11:04:15 - Il-2
Frames: 4288 - Time: 60000ms - Avg: 71.467 - Min: 59 - Max: 80

2020-12-18 11:10:09 - Il-2
Frames: 4301 - Time: 60000ms - Avg: 71.683 - Min: 59 - Max: 81

 

Thank you for running the CPU and GPU test.

Your CPU test is quite good and aligned with other Zen3 CPUs. BTW, did you use some overclock or PBO or just stock?

 

But in the GPU something weird is happenning. You have a 2080Ti but you only achieved 71fps, while I with a lower CPU score and 1080Ti obtain 94fps at 4K. And another 1080Ti reached 90.7. I really don´t know this.

Did you used a 4K monitor or did you use DSR? Could you bench your GPU with other public benchmarks?

 

 

 

Edited by chiliwili69
Link to comment
Share on other sites

15 hours ago, hellfirelx said:

When it was stable in multiple core stress test where it ran 4550 all cores, it failed in single core, where it always hit 4850 on the 2 prefered cores

Many many thanks for these extra test before your new 3080.

 

I think that many tests (like Prime95 with 4 threads) stress a lot the CPU to limits that will never be experienced by IL-2. So we are leaving on the table some extra fps just for IL-2.

 

You have a system almost identical to mine: B550, 5600X, 240mm AIO, RAM 3600-CL16, 1080Ti and Index!!.

Our CPU tests are very similar but your VR tests (both) are much much higher than mine. 

Your higher score in the VRTest1 could be due to the +200PBO and the hihger score in VRTest2 could be due to OC in the GPU?

Link to comment
Share on other sites

WheelwrightPL
2 hours ago, chiliwili69 said:

 

Thank you for running the CPU and GPU test.

Your CPU test is quite good and aligned with other Zen3 CPUs. BTW, did you use some overclock or PBO or just stock?

 

But in the GPU something weird is happenning. You have a 2080Ti but you only achieved 71fps, while I with a lower CPU score and 1080Ti obtain 94fps at 4K. And another 1080Ti reached 90.7. I really don´t know this.

Did you used a 4K monitor or did you use DSR? Could you bench your GPU with other public benchmarks?

 

 

Thank you for looking into this.

I just ran a free PassMark Performance Test 10.0 and my 3D MARK score for 2080Ti was 27606. What is the score for 1080Ti ?

Could the discrepancy you mentioned be caused by MSAA Antialiasing set to x8 ? (as per your instructions).

I think x8 setting should not be used because it is a recent addition to IL2-graphics-settings, and therefore we cannot compare thus obtained GPU scores with older scores which only had x4.

Maybe that is what is happening with those 1080Ti scores which seemingly exceed my 2080Ti ?

 

My PassMark Performance Test 10.0 3D MARK score looks fine:

 

 

vdcs.jpg

Link to comment
Share on other sites

Just for checking the new version 4.504. It gives exactly the same performance in the CPU and GPU Test.

So all previous test in version 4.503 are equally valid and will be mixed with 4.504

I didn´t change anything else in my computer from tests in 4.503.

 

CPU test v4.504
Frames: 7008 - Time: 60000ms - Avg: 116.800 - Min: 105 - Max: 154
Frames: 7050 - Time: 60000ms - Avg: 117.500 - Min: 103 - Max: 156
Frames: 7013 - Time: 60000ms - Avg: 116.883 - Min: 103 - Max: 152
Frames: 7126 - Time: 60000ms - Avg: 118.767 - Min: 105 - Max: 163

 

GPU Test v4.504
Frames: 5785 - Time: 60000ms - Avg: 96.417 - Min: 79 - Max: 114
Frames: 5797 - Time: 60000ms - Avg: 96.617 - Min: 80 - Max: 114
Frames: 5673 - Time: 60000ms - Avg: 94.550 - Min: 76 - Max: 113
Frames: 5722 - Time: 60000ms - Avg: 95.367 - Min: 79 - Max: 114
Frames: 5728 - Time: 60000ms - Avg: 95.467 - Min: 78 - Max: 112

  • Upvote 1
Link to comment
Share on other sites

8 hours ago, chiliwili69 said:

You have a system almost identical to mine: B550, 5600X, 240mm AIO, RAM 3600-CL16, 1080Ti and Index!!.

Our CPU tests are very similar but your VR tests (both) are much much higher than mine. 

Your higher score in the VRTest1 could be due to the +200PBO and the hihger score in VRTest2 could be due to OC in the GPU?

 

Hi,

Well all I know is your 1080ti card has this base characteristics :

OC Mode - GPU Boost Clock : 1708 MHz , GPU Base Clock : 1594 MHz Gaming Mode (Default) - GPU Boost Clock : 1683 MHz , GPU Base Clock : 1569 MHz

Mine runs the test with a steady 1974  Mhz, so well yes I guess that could be the explanation for both VR Testresults. Because if the PBO+200 would mean anything, wouldn't it bring my CPU Test over yours ? It only reaches this when I open the Power Limit of the GPU above 100%. Else it is power limited and the clock is lower. Now it is only voltage limited, but that's seems to be the normal thing with the nvidia boost algorithm. You can't damage anything, the limits will either bring the clock down or you get errors.

 

Link to comment
Share on other sites

SCG_Fenris_Wolf

@hellfirelx 

Please note, I commented earlier that I found out that if you use PBO or raise the stock clock, Zen3 performs worse.

 

To get the best performance, disable PBO, disable GameMode, keep stock clock. Let the AMD drivers handle it by themselves.

 

It performs best out of the box in game scenarios.

 

If you want to manually tune performance, look at the RAM. :good:

Link to comment
Share on other sites

On 11/23/2020 at 11:41 AM, LF_Gallahad said:

Okay! first OC test CPU

 

I used +100 PBO augment plus 1.100 Vcore SOC and 0.950 VDDP

 

Corsair H110i PRO XT (Maximum performance enabled) Peak of 63 ºC

 

Hey, I have a system very similar to yours: Same Mobo+CPU+RAM speed (I just followed proven hardware) and also AIO liquid cooling.

 

I would like to reach same performance as you. I understand you used BIOS for that, right?  From were did you learn to OC? Any recomendation?

 

You say you used 0.95 VDDP, but in my BIOS the maximum is +0.7V.

Link to comment
Share on other sites

17 minutes ago, chiliwili69 said:

 

Hey, I have a system very similar to yours: Same Mobo+CPU+RAM speed (I just followed proven hardware) and also AIO liquid cooling.

 

I would like to reach same performance as you. I understand you used BIOS for that, right?  From were did you learn to OC? Any recomendation?

 

You say you used 0.95 VDDP, but in my BIOS the maximum is +0.7V.

Hello chilli by trial and error.

 

I was in f11d Gigabyte bios for 550, but there are A LOT of problems right now with mutiple bios and even, they have gone up to f12a in our MOBO. Maybe there are some settings being problematic. We need to wait until Gigabyte finishes his business.

 

My guess is that I was able to do that in that BIOS, but also there were some WHEA errors in the long run. I decided to update and I don't have that one anymore. The Gigabyte bios are a completly mess rn, but looks like other brands have this problem too as MSI is going to skin 1.1.8.0 AGESA.

Link to comment
Share on other sites

7 minutes ago, LF_Gallahad said:

Hello chilli by trial and error.

 

I was in f11d Gigabyte bios for 550, but there are A LOT of problems right now with mutiple bios and even, they have gone up to f12a in our MOBO. Maybe there are some settings being problematic. We need to wait until Gigabyte finishes his business.

 

Ok thanks.

 

But just to checking, my current BIOS is F11n. It is the latest from Gigabyte webpage https://www.gigabyte.com/Motherboard/B550-AORUS-PRO-rev-10/support#support-dl-bios

 

Why do you say f12a?

BIOS.thumb.png.3e75317b7982c6191b9ee6213434479f.png

Link to comment
Share on other sites

1 minute ago, chiliwili69 said:

 

Ok thanks.

 

But just to checking, my current BIOS is F11n. It is the latest from Gigabyte webpage https://www.gigabyte.com/Motherboard/B550-AORUS-PRO-rev-10/support#support-dl-bios

 

Why do you say f12a?

BIOS.thumb.png.3e75317b7982c6191b9ee6213434479f.png

Yes, I used the beta BIOS from a page as I had many BSOD https://www.tweaktownforum.com/forum/tech-support-from-vendors/gigabyte/28656-gigabyte-latest-beta-bios?p=976054#post976054

I am testing F10 right now... and waiting for a final F11.

Link to comment
Share on other sites

Well, this is a lot of info and tweaking for just few extra fps. Perhaps I will leave it for now until it gets easier and safer.

In the Intel it was much simpler and measurable, but as Fenris say, the default stock settings seems to give the best performance. And this is nice BTW.

 

Unless there is a simple and widely accepted trick for the 5600X and B550 Mobos, I will run with stock settings (only XMP enabled) and have fun with IL-2.

 

Link to comment
Share on other sites

SCG_Fenris_Wolf
On 12/19/2020 at 8:49 AM, chiliwili69 said:

 

This is the SS you applied for the VRTest1, but this is alsmot 19.5 million pixels (3324x2824x2), so I will take it as the VRTest2.

It is indeed a very good result for the 3090. You reached the 75fps limit because your Pimax 8KX runs at 75Hz.

 

In your second test you went quite beyond the 19.5 million pixels. you did it at SS 166% (4788x4068x2) which is 38.9 million pixels, and you got 73fps avg. That´s really amazing!!!

That 3090 is insane.

 

Is VRTest 2 18.6M pixel now, or 19.5M pixel?

:S

 

p.s. True rgd MSI BIOS, I run .151beta to get no WHEA errors. FCLK max is 1900 on it.

 

 

 

Guys, I noticed that SteamVR doesn't change my resolution properly when selecting it individually "per application" in the video tab. It always stays at 100% when checking in-game visa fpsVR. The global slider properly changes it. New bug in SteamVR?

 

Did you also noticed this or is my SteamVR client kaputt?

Edited by SCG_Fenris_Wolf
  • Like 1
Link to comment
Share on other sites

7 hours ago, SCG_Fenris_Wolf said:

 

Is VRTest 2 18.6M pixel now, or 19.5M pixel?

:S

 

p.s. True rgd MSI BIOS, I run .151beta to get no WHEA errors. FCLK max is 1900 on it.

 

 

 

Guys, I noticed that SteamVR doesn't change my resolution properly when selecting it individually "per application" in the video tab. It always stays at 100% when checking in-game visa fpsVR. The global slider properly changes it. New bug in SteamVR?

 

Did you also noticed this or is my SteamVR client kaputt?


Since they updated the SteamVR app a couple of days ago I've had it crash twice on me, once mid-game.

Link to comment
Share on other sites

8 hours ago, SCG_Fenris_Wolf said:

Is VRTest 2 18.6M pixel now, or 19.5M pixel?

:S

 

The VR test2 is 19.5Mpixel. just to be on pair with the 100%SS of the G2.

8 hours ago, SCG_Fenris_Wolf said:

Guys, I noticed that SteamVR doesn't change my resolution properly when selecting it individually "per application" in the video tab. It always stays at 100% when checking in-game visa fpsVR. The global slider properly changes it.

 

I never used per application setting (which remains at 100%). I always used the global (General or Video). This is because I mostly only play one VR game. (IL-2).

Link to comment
Share on other sites

13 hours ago, SCG_Fenris_Wolf said:

Guys, I noticed that SteamVR doesn't change my resolution properly when selecting it individually "per application" in the video tab. It always stays at 100% when checking in-game visa fpsVR. The global slider properly changes it. New bug in SteamVR?

 

Isn't it like that the global slider is your baseline and the per app slider is relative to that? So if you set the global slider to say 70% that resolution is going to be your 100% on the per app slider.

Link to comment
Share on other sites

SCG_Fenris_Wolf
3 hours ago, HunDread said:

 

Isn't it like that the global slider is your baseline and the per app slider is relative to that? So if you set the global slider to say 70% that resolution is going to be your 100% on the per app slider.

Yes, that is how it should be. But the app slider does nothing, it only shows change in SteamVR. But even a restart doesn't update the actual setting used in-game. It shows xy% in SteamVR, but doesn't multiply the global one anymore if you load up the game.

 

It may be the reason some guys had bad benchmarking results

Edited by SCG_Fenris_Wolf
Link to comment
Share on other sites

thermoregulator
3 hours ago, SCG_Fenris_Wolf said:

Yes, that is how it should be. But the app slider does nothing, it only shows change in SteamVR. But even a restart doesn't update the actual setting used in-game. It shows xy% in SteamVR, but doesn't multiply the global one anymore if you load up the game.

 

It may be the reason some guys had bad benchmarking results

I can confirm that, per app settings slider does nothing on my end too.

 

Btw, a little OT, I am still struggling to find good settings for 5900x, rtx 3090 and reverb g2. With reverb g1, I could just max. everything out and was able achieve stable 90 fps. With g2, 100% SteamVR SS, it isn't that case, so I have to lower some settings or turn on the reprojection. I am not really happy with the reprojection, because after the years, I finally have a CPU, that is able to hold 90 FPS in VR, and with reprojection enabled, it is useless. Also, I have a feeling that with reprojection on auto, it is harder to keep constant 90 FPS than with reprojection off. Which settings dfdo you guys prefer to tone down with the g2, steam SS value, or some in game graphics settings? Do you prefer reprojection off, auto, or forced on? Have anyone tried legacy reprojection mode?

Link to comment
Share on other sites

12 hours ago, thermoregulator said:

Btw, a little OT, I am still struggling to find good settings for 5900x, rtx 3090 and reverb g2.

 

This is not OT, all you ask is about performance.

You have a pretty nice CPU with RAM at 3800CL16 and Infinity Fabric at 1900MHz. So, the CPU should not be bottleneck even with good settings.

The same with the 3090, this card is a beast and it can handle those 100%SS with no problem.

 

The benchmark settings are a bit weird, but they are like that because its mission is to stress just CPU or just GPU in monitor. The VR tests are just a bit of a mix, but the VRTest1 is looking to see the CPU limit and the VRTest2 (if VR test1 is fine) is looking to stress more the GPU.

 

I was studying the impact of each graphics settings here and the influence of in-game FXAAVsMSAA here.

 

You can leave the 100%SS settings in SteamVR (be sure that the numbers sum up 19.5Mpixels). You have paid full price for G2 and 3090, so get their juice.

Then put mirrors, cannopy reflections and distant building to off.

Put AA to FXAAx2 or no AA at all. At the G2 resolution and 100%SS you will not need them at all.

Everything else can be maxed out.

Then let me know if you enter in reprojection mode in complex SP missions. (I don´t know about MP). You can use fpsVR while you play just to know how good are your frametimes in each situtation. The fps in-game counter is OK but it is not giving you info about how big is your margin to enter in reprojection for both CPU and GPU.

 

Please, try to run the CPU test and GPU test, so we have a better picture of your system.

 

  • Like 1
Link to comment
Share on other sites

thermoregulator
3 hours ago, chiliwili69 said:

 

 

 

Please, try to run the CPU test and GPU test, so we have a better picture of your system.

 

Will try later today, I have just have to solve "60 Hz lock problem" on my system.

Link to comment
Share on other sites

thermoregulator
5 hours ago, chiliwili69 said:

 

 

 

Please, try to run the CPU test and GPU test, so we have a better picture of your system.

 

I am sorry, I tried again and I just couldn't do that. In 2D, I am always locked on 60 Hz in this game, even though v-sync si off. Every other game I have works OK, but IL-2 is locked on 60 in fraps (v-sync off on both global and IL-2 profile). So I can do just VR Benchmark (Which I have done before with G1).

Link to comment
Share on other sites

1 hour ago, thermoregulator said:

I am sorry, I tried again and I just couldn't do that. In 2D, I am always locked on 60 Hz in this game, even though v-sync si off. Every other game I have works OK, but IL-2 is locked on 60 in fraps (v-sync off on both global and IL-2 profile). So I can do just VR Benchmark (Which I have done before with G1).

Thanks for trying.

I have been also trying to guess what it could be.

First at all, check that your monitor is not equiped with G-Sync or Freesync technologies. (mire has G-sync and I have to switch it off for tests).

Secondly, try to eliminate the profile of IL-2 from the NVIDIA panel, so all go with global setting.

There is also the Adaptive Sync tech from NVIDIA which uses DisplayPort. Look if there is more features related to sync in the all the options of all sections of NVIDIA panel settings.

What monitor brand you do have? Do you use DP cable for the monitor?

Link to comment
Share on other sites

SCG_Fenris_Wolf

I'd reinstall Nvidia Drivers cleanly, reboot, check WMR software for 90Hz setting, reinstall SteamVR cleanly (delete profiles and settings manually too) and disable any G-Sync or Freesync tech. 

 

 

I can't bench my newly arrived RTX 3090 right now because the 750W PSU can't handle it together with the 5900X. I got a new one (Seasonic Prime GX 850) on order, but I doubt it comes before X-Mas. Shouldn't have selected next-day express shipping, they just skipped me with the stock that came in today :)  Damn

Edited by SCG_Fenris_Wolf
Link to comment
Share on other sites

5 hours ago, thermoregulator said:

but IL-2 is locked on 60 in fraps

 

Just a clarification. When you run fraps in monitor you will see the fps in yellow over the IL-2 game. During the menu and mission selection the counter is loack at 60fps, but when you launch the mission the fps go then above 60.

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...