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

[1Sascha]

11 minutes ago, dburne said:

I ramp all my fans up to max - I use earbuds and they isolate the noise pretty good for me. I try and keep my rig as cool as I can whilst gaming.

 

Yeah.. I was just about to fire up Afterburner to see what the max RPM for the GPU fans is. If 2000 is already their max speed, then there's not much I can do, I guess. All my case-fans do ramp up even when I start SteamVR, I set them up pretty cautiously in my BIOS. But since my case lacked a few mm to install my AiO in the top, I had to put it in the front and so now all the fresh air I suck into the case goes through the rad first. So not the best ventilation around the GPU...

 

Oh, well... I did just reduce IL-2 SS from 136 to 110, that should make a difference in GPU load, I suppose.

 

EDIT: Holy crap! 100% fan speed on these is ~3500 RPM... and then the card sounds like a leaf-blower.. ? Custom fan-curve: Here I come!

 

 

S.

 

Edited May 22, 2022 by 1Sascha

[chiliwili69]

6 hours ago, 1Sascha said:

Guess my PC is now a lot less bottle-necked by its GPU - which was what I was hoping for

 

Thank you for running again the benchmark.

As you can see now your CPU performance is good, at not bottlenecked by GPU. (this is the intend of the CPU test)

 

6 hours ago, 1Sascha said:

but 136% SS

 

Using a Rift-S for the VR-Test1 you should use 184% SS in SteamVR. In terms of "Pixel Density" it is 1.36. (Square_root of 1.84 is 1.36).

For the Oculus devices the instructions are given in terms of PD since you can set that in the Oculus software.

 

[1Sascha]

39 minutes ago, chiliwili69 said:

Using a Rift-S for the VR-Test1 you should use 184% SS in SteamVR. In terms of "Pixel Density" it is 1.36. (Square_root of 1.84 is 1.36).

For the Oculus devices the instructions are given in terms of PD since you can set that in the Oculus software.

Oops... my bad. I'll do that ASAP, but I already dread the results ... it's going to be laaaaaggy I fear... ?

 

S.

[chiliwili69]

10 hours ago, 1Sascha said:

it's going to be laaaaaggy I fear...

Don´t worry about that. The VR tests in this benchmark were designed having 9.5 million pixels (VR1 test) and 19.5 million pixels (VR2 test) which correspond to the 50% and 100% SS of the G2.

The Rift-S at 100% SS has 5.8 Million pixels at 80Hz.  The VR1 test for the Rift-S uses 184% SS (10.7 million at 80Hz is the same than 9.5million at 90Hz) in order to compare it with other devices at a common base of 90Hz. So, it is just a benchmark number, only useful to compare it with your peers (Rift-S users).

 

In normal use of Rift-S, 100% or 110% is quite enough to have a decent view. You can use Afterburner trendlines for GPU load and compare it with diferent settings of IL-2.

Avoid MSAA and extreme/high clouds to don´t overload GPU. Your 2070S is well suited for a headset like Rift-S.

[1Sascha]

On 5/23/2022 at 9:14 AM, chiliwili69 said:

Avoid MSAA and extreme/high clouds to don´t overload GPU. Your 2070S is well suited for a headset like Rift-S.

I already switched down to 2x FXAA and slightly reduced some other stuff to get a smoother performance in IL-2. Like I stated earlier: My current game-settings, even with 136% SS in SteamVR, "only" pushed the card to around 195 W or so... still ~20W to go before it hits its power-limit.. ?

 

 

EDIT: Already updated my original report, but here are the results again with SteamVR set to 184%

 

2022-05-23 11:27:51 - Il-2
Frames: 2542 - Time: 60000ms - Avg: 42.367 - Min: 38 - Max: 62

 

I guess this means I probably shouldn't run 184% when I'm actually playing the game. As far as I understand, it is recommended to try and get your FPS as close as possible to your set's max refresh-rate? Sooo.... 80, in the case of the Rift-S?

 

I mean... the GPU didn't get pushed as hard as I had feared, but still above 190W and that's with the super-reduced in-game settings the test calls for. So, I don't think I'll stay at 184% for gameplay, even though the picture looked a bit crisper to me in the menus.

 

Well... at least my CPU is still pretty bored by it all ... ?

 

 

 

 

EDIT: Sorry for going slightly OT, but will the FRAPS benchmark function give me any meaningful numbers when it comes to VR and actual gameplay? Like I said, I'm trying to get as close to my Rift's max refresh rate as I can and I'm willing to sacrifice quite a bit of eye-candy, but I don't know how to reliably measure and compare FPS to confirm how effective or ineffective my changes to the in-game settings are.

 

 

 

S.

Edited May 27, 2022 by 1Sascha

[chiliwili69]

9 hours ago, 1Sascha said:

2022-05-23 11:27:51 - Il-2
Frames: 2542 - Time: 60000ms - Avg: 42.367 - Min: 38 - Max: 62

 

I guess this means I probably shouldn't run 184% when I'm actually playing the game. As far as I understand, it is recommended to try and get your FPS as close as possible to your set's max refresh-rate? Sooo.... 80, in the case of the Rift-S?

 

Uhmm, your Avg for this card and CPU should be at least around 60 fps comparing it to another tests of Rift-S in the table @SAG (it was with previous version 4.701) but with the same card than you (2070S).

 

So I am not sure if this 18fps difference is only because the change in version.

Technically your 2070S performs very well with the GPU public benchmarks.

 

I don´t know the tricks for the Rift-S (OpenComposite, any other thing? ) but for the same number of pixels it performs worse than other headsets like G1 or G2 or Index.

 

Every VR user has to take one crucial decission: To use MotionSmoothing (called ASW in Oculus) or not.

If you use ASW your target is 40fps (the other 40fps are invented based on trajectories), if you don´t use ASW your target is 80fps.

Some people live happy with that, other don´t like it. I personally prefer not to use MotionSmoothing, but it depends on every particular case.

 

I believe the ASW 2.0 is quite good, so try it and if you like it you will be able to raise quite a lot your settings.

 

The fpsVR app (cost 4 dollars) is quite good to see CPU and GPU performance while you are gaming. So you can see the influence of each settings.

 

 

 

 

[1Sascha]

I just reverted to my old IL-2 settings and 125 % SS in SteamVR, ran a QM with 8 fighters and 4 bombers and did some typical stuff like fighting, shooting planes, low level flight and landing all while running the Fraps benchmark.

 

2022-05-23 21:55:54 - Il-2
Frames: 3406 - Time: 60000ms - Avg: 56.767 - Min: 39 - Max: 80

 

It didn't feel stuttery to me and I didn't get noticeable glitches, plus the CPU was hovering around 75% usage/100W and the GPU only reached 76° max, 196W and 96% usage.

 

Speaking of glitches: I had those before (wings on other planes looked like they were flapping) and then I switched ASW off in the debug-tool. Flapping wings disappeared right away, so I'm not sure I'd want that thing back on... I think I might try to reduce in-game settings a bit more and switch ASW back on just to see if it'll make the game look better/smoother. Man, this is all complicated... ?

 

 

S.

[dburne]

10 minutes ago, 1Sascha said:

I just reverted to my old IL-2 settings and 125 % SS in SteamVR, ran a QM with 8 fighters and 4 bombers and did some typical stuff like fighting, shooting planes, low level flight and landing all while running the Fraps benchmark.

 

2022-05-23 21:55:54 - Il-2
Frames: 3406 - Time: 60000ms - Avg: 56.767 - Min: 39 - Max: 80

 

It didn't feel stuttery to me and I didn't get noticeable glitches, plus the CPU was hovering around 75% usage/100W and the GPU only reached 76° max, 196W and 96% usage.

 

Speaking of glitches: I had those before (wings on other planes looked like they were flapping) and then I switched ASW off in the debug-tool. Flapping wings disappeared right away, so I'm not sure I'd want that thing back on... I think I might try to reduce in-game settings a bit more and switch ASW back on just to see if it'll make the game look better/smoother. Man, this is all complicated... ?

 

 

S.

 

Main thing to shoot for is how it does for you in the game itself.

 

[1Sascha]

52 minutes ago, dburne said:

Main thing to shoot for is how it does for you in the game itself.

Well.. that's what I'm trying to optimize. I'm probably running into limitations re the Rift-S and/or my card, but I did notice that objects (and icons/plane markers for that matter) seem to look rather fuzzy at around 3 to 4 km distance. Maybe it's my eyes being unable to focus or it's technical limitations (lack of pixels, etc), but I think I'll be tinkering a while longer. Plus somehow ASW was enabled/set to Auto again when I just double-checked. I wonder if SteamVR can override that setting in the Oculus SW somehow?

 

S.

[dburne]

14 minutes ago, 1Sascha said:

Well.. that's what I'm trying to optimize. I'm probably running into limitations re the Rift-S and/or my card, but I did notice that objects (and icons/plane markers for that matter) seem to look rather fuzzy at around 3 to 4 km distance. Maybe it's my eyes being unable to focus or it's technical limitations (lack of pixels, etc), but I think I'll be tinkering a while longer. Plus somehow ASW was enabled/set to Auto again when I just double-checked. I wonder if SteamVR can override that setting in the Oculus SW somehow?

 

S.

 

No planes at a distance are going to be somewhat fuzzy in VR. Grab the Oculus Tray Tool  you can use it to set your ASW preferences per game along with other things.

https://www.apollyonvr.com/oculus-tray-tool

[1Sascha]

Good call re VR Tray Tool. Reset SteamVR's slider to 100%, then made a simple VRTT-profile for IL-2 just setting SS to 1.2 and disabling ASW. Once I actually managed to open everything in the correct order and it did give me the "launch detected" audio-cue, the game ran quite a bit better for me. I'm not sure the profile is indeed what changed things, but since I didn't change anything else I don't see what else could've caused this: With the profile running correctly, my CPU usage and power-draw seem to have gone down quite a bit. And with my current in-game settings I'm now pretty solidly into 70-80FPS territory. Which is probably as good as it'll get without sacrificing too much eye-candy.

 

S.

[chiliwili69]

49 minutes ago, 1Sascha said:

my CPU usage

I suposse you refer to GPU usage, since it was your bottleneck.

 

Probably using SS in SteamVR and DEbug tool PD at the same time could lead to duplicate the total supersampling. Perhaps The tray-tool has reset all to the right values and you are now only applying SS in one place via tray tool.

 

The ASW On is the responsible of those flapping wings, and this happen independently of the IL-2 settings or SS. So if you don´t like better to turn ASW Off.

 

Gald to hear you a re now in the 70-80 fps region. One of the good things of the Rift-S (like Index) is runing at 80Hz (no 90Hz), so it requires a bit less resources from CPU and GPU.

 

You might want to run the VRtest1 again just to see if it improved.

 

Also, people used OpenComposite to have better performance with Oculus devices.

Edited May 25, 2022 by chiliwili69

[1Sascha]

2 hours ago, chiliwili69 said:

I suposse you refer to GPU usage, since it was your bottleneck.

Nope... I meant CPU... which was quite a bit higher before I managed to switch off ASW reliably and to apply the VRTT profile. About 20 - 25W higher max power-draw and also a bit more toasty. Not sure what happened and it could be coincidence or the fact that I'm only using HWMonitor and its "max" readout... not constantly monitoring temps/power draw while in the cockpit.

 

3 hours ago, chiliwili69 said:

Probably using SS in SteamVR and DEbug tool PD at the same time could lead to duplicate the total supersampling. Perhaps The tray-tool has reset all to the right values and you are now only applying SS in one place via tray tool.

 

Yeah... I did double-check that SS in SteamVR was reset to default/100 before using the Tray Tool. But some of this stuff seems pretty counter-intuitive and some settings don't seem to stick for me... or they didn't before I used the TrayTool. ASW turned off in the Debug-Tool for example would be set to Auto again when I checked it again before starting the next IL-2 session. Not sure which app is "top dog", because it seems to me that some settings I make in one app are simply overwritten by the other. ?

 

 

3 hours ago, chiliwili69 said:

You might want to run the VRtest1 again just to see if it improved.

 

I'll definitely do that.

 

S.

[dburne]

4 hours ago, 1Sascha said:

Good call re VR Tray Tool. Reset SteamVR's slider to 100%, then made a simple VRTT-profile for IL-2 just setting SS to 1.2 and disabling ASW. Once I actually managed to open everything in the correct order and it did give me the "launch detected" audio-cue, the game ran quite a bit better for me. I'm not sure the profile is indeed what changed things, but since I didn't change anything else I don't see what else could've caused this: With the profile running correctly, my CPU usage and power-draw seem to have gone down quite a bit. And with my current in-game settings I'm now pretty solidly into 70-80FPS territory. Which is probably as good as it'll get without sacrificing too much eye-candy.

 

S.

 

Good deal have some fun now!!

[1Sascha]

Still got minor trouble with SS though. Tray Tool will detect game-launch now and I'm using its overlay to check performance and SS setting - and to switch ASW on the fly, which is pretty handy.

 

Thing is that SS settings I set in the tool don't necessarily show up as active once the game runs. For example: Setting 1.2 SS in the TT's IL-2 profile will still result in pixel density 1.0/1.0 in the overlay once the game runs. Only when I up SS in the TT to, err.. 1.5 IIRC did I see a value higher than 1.0/1.0 in-game (1.23 in this case IIRC). I guess that's why TT recommends using the overlay to confirm what kind of SS you actually get with TT's multipliers.

Also: In one of the first screens of the TT in-VR overlay ("Layer 0 info"), there's a "High Quality" entry (second item from the top, IIRC). It's always set to "OFF" and I have no idea how to turn it on (if that's even possible?). I did set the Oculus SW to "emphasize quality", but I don't know if that has anything to do with it.

 

Anyone know what I'm talking about here?

 

Plus: What's the preferred method of setting SS? From what I understand, I can do it in the TT, but that seems rather imprecise, since I have to use fixed values and never know what the actual SS will be in-game. 1.23 actual seems a bit too demanding for my card and all the TT-settings I've tried so far either gave me 1.0 or 1.23... nothing in-between. So would it be better to reset TT to 1.0 and use SteamVR to set IL-2's SS?

 

S.

Edited May 27, 2022 by 1Sascha

[RossMarBow]

On 4/23/2022 at 10:01 PM, shirazjohn said:

I've been running with inf fabric at 2000 mhz with no problem at all, i don't know if the chipset makes a difference x570 v b550. I have been running at 4000 mhz for probably a year and its been rock solid.

 

This is the memory I have I'm no expert just bought what i thought was right at the time.

https://www.overclockers.co.uk/team-group-8pack-ripped-edition-16gb-2x8gb-ddr4-pc4-28800c14-3600mhz-dual-channel-kit-black-my-001-8p.html

how are you running IF at 2000 mhz 

 

is it actually linpack stable?

can you share your zen timings?

[RAAF492SQNOz_Steve]

13 hours ago, brahguevara said:

how are you running IF at 2000 mhz 

 

is it actually linpack stable?

can you share your zen timings?

Running IF at 2000 Mhz etc can be very dependent on the CPU lottery and playing with Zen timings may make little difference as has just been demonstrated to me.

Have just upgraded/ sidegraded from a 5800x to a 5800x3D and had to reduce my memory and IF speeds to obtain stability.
Formerly could run RAM @3933 and IF at 1967Mhz with 5800x. The 5800x3D best (stable) setting is RAM 3867MHz and IF of 1933 MHz.

 

All that has changed is the CPU.   No RAM timing changes alter that issue and memory etc was not changed during the upgrade.
The memory itself was rated for 4000 MHz operation (and has run at that in the past although during really hot weather became unstable so backed it off a bit).

 

Results of upgrade for IL2.

Noticed a 8% instant jump with Varjo Aero average IL2 framerates after the upgrade to the 3D with the SVander benchmark.
Was very much due to low fps figure increasing rather than being able to run higher frame rates.
Aero was running 35 ppd resolution setting and GPU was a RTX 3090.

 

FYI,
Can run lower timings but the memory test tools I have reported a deterioration in data throughput so have backed tRFC & tRC settings off somewhat. 

 

 

 

[shirazjohn]

On 5/27/2022 at 11:47 AM, brahguevara said:

how are you running IF at 2000 mhz 

 

is it actually linpack stable?

can you share your zen timings?

 

I used the following settings with my old 5600x and ran stability tests which showed no error's but must confess that i haven't run any stability tests with the x3d yet which is something i must do after what the previous post remarked about reducing his settings after installing the x3d. 

 

[chiliwili69]

15 hours ago, RAAF492SQNOz_Steve said:

Noticed a 8% instant jump with Varjo Aero average IL2 framerates after the upgrade to the 3D with the SVander benchmark.

Quite nice. It would be interering to have some Aero numbers in the SYNvander table, just to track improvements.

On 5/27/2022 at 12:47 PM, brahguevara said:

how are you running IF at 2000 mhz

 B550 is worse than X570.  And brand also matters.

You only need to check the proven and guaranteed XMP profiles of Gskill for every mobo:

 

For my B550 Mobo best I can get is 3600MhZ with 16-16-16-36:

 

https://www.gskill.com/configurator?page=1&cls=1529635169&manufacturer=1524715126&chipset=1603951212&model=1603952231&adSearch2=Capacity§32GB (16GBx2),Tested_Speed§3600 MT/s,

 

But for TFU Gaming X570 Plus you have 4000MHz with 16-16-16-36:

 

https://www.gskill.com/configurator?page=1&cls=1529635169&manufacturer=1524725352&chipset=1603955103&model=1603956523&adSearch2=Tested_Speed§4000 MT/s,

 

And if you go to the EVGA X570 Dark you have 4000Mhz 14-15-15-35:

 

https://www.gskill.com/configurator?page=1&cls=1529635169&manufacturer=1524715213&chipset=1634872429&model=1634872781&adSearch2=Tested_Speed§4000 MT/s,Tested_Latency§14-15-15-35,

 

So, I made a mistake when I pick a Gigabyte B550. Next time I will look at supported memories first to decide the Mobo.

[RossMarBow]

11 hours ago, chiliwili69 said:

Quite nice. It would be interering to have some Aero numbers in the SYNvander table, just to track improvements.

 B550 is worse than X570.  And brand also matters.

You only need to check the proven and guaranteed XMP profiles of Gskill for every mobo:

 

For my B550 Mobo best I can get is 3600MhZ with 16-16-16-36:

 

https://www.gskill.com/configurator?page=1&cls=1529635169&manufacturer=1524715126&chipset=1603951212&model=1603952231&adSearch2=Capacity§32GB (16GBx2),Tested_Speed§3600 MT/s,

 

But for TFU Gaming X570 Plus you have 4000MHz with 16-16-16-36:

 

https://www.gskill.com/configurator?page=1&cls=1529635169&manufacturer=1524725352&chipset=1603955103&model=1603956523&adSearch2=Tested_Speed§4000 MT/s,

 

And if you go to the EVGA X570 Dark you have 4000Mhz 14-15-15-35:

 

https://www.gskill.com/configurator?page=1&cls=1529635169&manufacturer=1524715213&chipset=1634872429&model=1634872781&adSearch2=Tested_Speed§4000 MT/s,Tested_Latency§14-15-15-35,

 

So, I made a mistake when I pick a Gigabyte B550. Next time I will look at supported memories first to decide the Mobo.

Just because it at can 4000 mhz
Doesn't mean you actually should most chips can only run a ~1900 IF whea free
so 3800 is the fastest you would run your ram anyway
but you could probably get better timings

 

so yea checking the memory list is a good idea

[Voyager]

@RAAF492SQNOz_Steve Moore's Law is Dead had a mention in one of the recent long form episode that, apparently Vcache isn't entirely ready for prime time, and tends to be squirrelier about voltage tolerances than we'd normally like for this. That's probably why they ended up disabling overclocking on it. 

 

On the other hand, I've found that memory timings seem to have little to no impacts on gaming performance, especially in VR. My guess is the cache is large enough that the CPU doesn't have to go to ram unexpectedly. 

 

Will be interesting to see what happens with Zen 4 Vcache

[shirazjohn]

Ran the Linpack stability test at 4000 mhz and had multiple whea errors even though I've had no ctd's , crashes, bsod's etc since installing the x3d and running at 4000mhz (about a month of daily use).

I had to drop down to 3800 mhz to get it stable after some testing between 3800mhz and 3600mhz there wasn't a great deal of difference in performance between the two in fact I had  better performance at 3600 mhz which I've now settled with.

 

memory at 3600 mhz 
2022-05-29 11:38:17 - Il-2
Frames: 7743 - Time: 60000ms - Avg: 129.050 - Min: 114 - Max: 182

2022-05-29 11:40:08 - Il-2
Frames: 7812 - Time: 60000ms - Avg: 130.200 - Min: 111 - Max: 188

2022-05-29 11:42:00 - Il-2
Frames: 7551 - Time: 60000ms - Avg: 125.850 - Min: 110 - Max: 170

Three run Avg = 128.36

 

memory at 3800 mhz
2022-05-29 10:51:28 - Il-2
Frames: 7630 - Time: 60000ms - Avg: 127.167 - Min: 109 - Max: 179

2022-05-29 10:53:20 - Il-2
Frames: 7518 - Time: 60000ms - Avg: 125.300 - Min: 110 - Max: 169

2022-05-29 10:55:15 - Il-2
Frames: 7470 - Time: 60000ms - Avg: 124.500 - Min: 108 - Max: 168
Three run Avg = 125.65

 

[shirazjohn]

Hi guys I ran another test today at 3200 mhz and I've come to the conclusion you have similar results whatever frequency your ram running at its the X3D larger cache that is doing all the work, maybe tighter timings and lower frequency are the way to go.

 

memory at 3200 mhz 

2022-05-30 17:49:18 - Il-2
Frames: 7563 - Time: 60000ms - Avg: 126.050 - Min: 110 - Max: 173

2022-05-30 17:51:05 - Il-2
Frames: 7797 - Time: 60000ms - Avg: 129.950 - Min: 114 - Max: 178

2022-05-30 17:52:57 - Il-2
Frames: 7973 - Time: 60000ms - Avg: 132.883 - Min: 115 - Max: 194

Average of 3 runs = 129.627

 

Edited May 30, 2022 by shirazjohn

[Voyager]

@chiliwili69

Vive Released a new beta version of their VP2 console that lets you disable the frame rate ratios: 2.1.21.0 Beta. You have to get it through Steam, which is weird and clunky, but it works.

 

It's under Settings > Video > FPS Preference.

 

I re-rant he VR tests with the new console and got the following results:

 

	VR Test 1 Vive Pro 2: (Ultra: 2200x2200)
	Min	Max	Avg
1	64	91	82.583
2	71	91	85.717
3	68	91	87.117
	67.66667	91	85.139
	VR Test 2 Vive Pro 2: (Ultra: 3148x3148)
	Min	Max	Avg
1	62	91	79.567
2	67	91	81.367
3	65	91	84.483
	64.66667	91	81.80566667
	VR Test 1 Vive Pro 2: (Extreme: 2184x2184)
	Min	Max	Avg
1	59	118	81.833
2	63	121	86.75
3	59	119	80.65
	60.33333	119.3333	83.07766667
	VR Test 2 Vive Pro 2: (Extreme: 3144x3144)
	Min	Max	Avg
1	59	101	77.333
2	66	101	83.233
3	63	102	82.267
	62.66667	101.3333	80.94433333

 

 

So no huge changes in the VR Test 1, but in the VR Test 2 it does seem to have a pretty significant improvement in average framerates. Also, the 120hz mods and 90hz modes seem to have roughly the same overall performance, though the 120 can push up to higher fps in some spots.

 

Addendum: Forgot to add the rest of the system config;

Configuration 1: Ryzen 5800X3D with 3080 Ti, and DDR4-3600
 Motherboard:    Gigabyte X570 Aorus Xtreme v 1.2
 Bios Version F36a

 CPU:        AMD Ryzen 7 5800X
 CPU Freq:    4450 MHz
 L3 cache:    2x16 MB
 Cores:        8
 Threads:    16

 RAM type:    DDR4
 RAM size:    128 GB (4x32)
 NB Freq:    1799.6 MHz
 RAM Freq:    1799.6
 RAM timings:    16-22-22-42
 
GPU:        EVGA 3080 TI FTW3, Core 1800Mhz, Memory 9501Mhz (Factory overclock)
 Driver Version 512.95

VR Headset 1: Vive Pro 2,  2.1.21.0 Beta

Game Version 4.705c

Edited June 5, 2022 by Voyager

[chiliwili69]

9 hours ago, Voyager said:

VR Test 2 Vive Pro 2: (Ultra: 3148x3148)   Min Max Avg 1 62 91 79.567 2 67 91 81.367 3 65 91 84.483   64.66667 91 81.80566667

Yeah, that beta version produces a more expectable results from the VP2 at high resolution. Now your performance is on pair with the G2. And both with great results.

 

I wonder how far the 12900K can go in VR (or just CPU or GPU test) with the latest versions of the game (which load more CPU and GPU). @DBCOOPER011?  @FoxbatRU?  @WallterScott?

[FoxbatRU]

4 часа назад, chiliwili69 сказал:

I wonder how far the 12900K can go in VR (or just CPU or GPU test) with the latest versions of the game (which load more CPU and GPU).@FoxbatRU? 

 

I have not played for a long time, but now I tried to test the processor.

 

CPU test

4.705c

Frames: 8937 - Time: 60000ms - Avg: 148.950 - Min: 126 - Max: 198

4.701b
Frames: 9095 - Time: 60000ms - Avg: 151.583 - Min: 137 - Max: 199

 

?‍♂️

 

 Motherboard:  Asus ROG STRIX Z690-A GAMING WIFI D4

 CPU:                 12900K
 CPU Freq:        5.2 GHz
 L3 cache:         30 Mb
 Cores:               8
 Threads:           16 (turned off E-cores )
 RAM type:        DDR4
 RAM size:         16 GB (2x8)
 NB Freq:           4164 MHz
 RAM Freq:        3860 MHz 
 RAM Latency:  17-17-17-34
 GPU:                 RTX 3080

 OS:                    Windows 10 (x64, 21H2)

 IL2:                   4.705c

 CPU Cooling:   NZXT Kraken X73

[chiliwili69]

4 hours ago, FoxbatRU said:

Avg: 148.950 - Min: 126 - Max: 198

Thank you very much for runnning again this CPU test, it clearly show that the Intel 12900K is still the king with the latest IL-2 versions.

That 12th gen of Intel is really a beast for IL-2.

[AceVenturi]

I'm looking at a GPU upgrade for il2. I only got vr with reverb g2. Currently have 1080 but have seen some 6700xt cheap. Equivalent Nvidia still $200+ more. Do the AMD cards still have a far worse performance in il2? Has anyone tried the dxvk mod to see if it improves things for AMD?

I have 5600x and 16gb 3000mHz Ram on b450 mb.

[Voyager]

On 6/14/2022 at 3:11 PM, AceVenturi said:

I'm looking at a GPU upgrade for il2. I only got vr with reverb g2. Currently have 1080 but have seen some 6700xt cheap. Equivalent Nvidia still $200+ more. Do the AMD cards still have a far worse performance in il2? Has anyone tried the dxvk mod to see if it improves things for AMD?

I have 5600x and 16gb 3000mHz Ram on b450 mb.

Last I checked they do, and I'm seeing people having issues with the new drivers. 

 

It may sort out; AMD is doing a big driver push to get ready for RDNA3, but right now it's still dodgy as anything. 

 

Probably best to look for a 3080 or 3080 12Gb for Reverb G2 VR. I need to check how much vRam I'm using in Il-2. I know other flight sims push 11gb in VR, but I think Il-2 is lighter. 

[gnat]

38 minutes ago, Voyager said:

Last I checked they do, and I'm seeing people having issues with the new drivers. 

 

It may sort out; AMD is doing a big driver push to get ready for RDNA3, but right now it's still dodgy as anything. 

 

Probably best to look for a 3080 or 3080 12Gb for Reverb G2 VR. I need to check how much vRam I'm using in Il-2. I know other flight sims push 11gb in VR, but I think Il-2 is lighter. 

Maybe hold off on the high-end 30-series GPUs since the 4090, 4080, and 4070 should start arriving in September.

[WallterScott]

05.06.2022 в 12:48, chiliwili69 сказал:

I wonder how far the 12900K can go in VR (or just CPU or GPU test) with the latest versions of the game (which load more CPU and GPU). @DBCOOPER011?  @FoxbatRU?  @WallterScott?

Hi! It's strange, but I have even more fps with the latest version of Il2.

 Motherboard:  Asus ROG Maximus Z690 Apex

 CPU:                 12900K
 CPU Freq:        5.5 for all core, 58 - 2 core GHz
 L3 cache:         30 Mb
 Cores:               8
 Threads:           16 
 RAM type:        DDR5
 RAM size:         32 GB (2x16)
 NB Freq:           4500 MHz
 RAM Freq:        6400 MHz 
 RAM Latency:  30-38-38-52
 GPU:                 RTX 3090

 OS:                    Windows 11

 IL2:                   4.706

 CPU Cooling:   water chiller

 

2022-06-25 20:50:25 - Il-2
Frames: 9479 - Time: 60000ms - Avg: 157.983 - Min: 140 - Max: 227

Edited June 25, 2022 by WallterScott

[chiliwili69]

On 6/25/2022 at 8:01 PM, WallterScott said:

IL2:                   4.706

 

Many thanks. This is interesting. Your fps are amazing. Perhaps this new version introduced some performance improvement. I will need to test it as well.

[Dustybritches]

I do not play bench marking but I do play IL2 in VR at these settings. Just for amusement I ran your bench mark at my game settings on my Pimax 5K super @ 120hz.              2022-07-24 18:58:59 - Il-2
Frames: 5357 - Time: 60000ms - Avg: 89.283 - Min: 78 - Max: 104 

 

FRAPSLOG.TXT FRAPSLOG.TXT

[chiliwili69]

11 hours ago, Dustybritches said:

Frames: 5357 - Time: 60000ms - Avg: 89.283 - Min: 78 - Max: 104 

 

That´s a good result.

The thing that surprises me more is that you get that with a 6900XT GPU. Those AMD cards were not giving good results in IL-2 VR.

 

Two questions:

What IL-2 setting did you used? (same than VR test1 or 2 or others)?

What resolution numbers (pixels) do you get when you set SteamVR SS% at 100% for you Pimax5KSuper?

[Dustybritches]

What IL-2 setting did you used? (same than VR test1 or 2 or others)? No because those settings favor the wider 384 bit memory bus architecture of the Nvidia GPUs. They overload the 256 bit memory subsystem of my card and significantly increase frame times. I use lower settings that do not impose this penalty and still look very nice at SS 3732X3172. The steam super sampling at 100% looks like pixelated crap and only utilizes 45% of my GPUs power budget gimping performance by keeping the graphics driver at a lower power state.

 

What resolution numbers (pixels) do you get when you set SteamVR SS% at 100% for you Pimax5KSuper?  Steam SS100%=3136X2576

[chiliwili69]

On 7/26/2022 at 6:55 AM, Dustybritches said:

I use lower settings

 

What settings do you low more to not bottleneck your 6900XT?

The VR test1 and VR test2 uses the lowest IL-2 settings except for the preset which is HIGH.

What settings did you use then?

[Dustybritches]

I should have looked closer as my settings are nearly the same sorry for the confusion. I guess it simply comes down to using steam SS to bring the cards power level to near maximum.

[Voyager]

@Dustybritches Out of curiosity, have you seen any impact from the latest driver release? Apparently it boosted performance in a number of older titles, and was wondering how it handled here? 

[Dustybritches]

@Voyager

Your referring to version 22.7.1? I don't have it installed still on 22.5.1. I haven't really noticed any great improvements between driver versions over the last 6 months really for IL2. I only play in VR though so not a broad database to go on there. The biggest improvements for me have been selling my 5900x and getting the 5800X3D along with cranking up SS in steam VR.

[chiliwili69]

On 8/3/2022 at 6:47 AM, Dustybritches said:

my settings are nearly the same

The reason why you got such a performance with the 6900XT is because the pre-settings you used were "LOW". This reduces considerably the bubble and number of objects to be render and your GPU is less bottlenecked then. If you run it with HIGH preset then the numbers should be diferent.