My man, you don’t have a clue how often I asked several of your fellow TH-camrs to talk about VR performance when looking and benching CPU’s and GPU’s. Who cares anymore if a system can poop out 450FPS or 400 FPS on a flat screen. This is soooo appreciated and the way you have approached this is second to none. This is the most clear explained and visualized bench I have ever seen and you got my sub for that. This is great man and I foresee greatness in your channel. Keep it up. Kuddo’s
Thanks, this clears up a lot for me. I could never really grasp how it works. I subbed, I love when people explain the technical side of iRacing. I'm a mechinical engineer myself, so I understand the car side of the sim but the computer/software side is some kind of black magic. Keep going, you deserve those 1000 subs.
Clear explanation 👍 The examples at the end were very interesting. I'll definitely be re-thinking having cockpit mirrors. And exploring ASW. Have never turned it on, but it looks to do a very good job of reducing frametimes below the 12ish ms I need for my 80 Hz Rift S.
Thank you for this video. Finally I understand why I'm seeing stutters in VR, I couldn't understand why I had a solid FPS but it didn't feel smooth. Your stutter example at the start of the video is exactly what I am seeing. Thank you also for explaining in simple terms how to interpret the results of your testing. I've got a new 7800X3D arriving today, so hopefully that will help reduce my stutters.
This is great. I'm quite competent with computers most of the time unless it's something really complicated like actually coding or whatever and I am NOT a professional by any means, I just use pcs a lot and am usually the one to host servers, create mod packs for games for my friends etc but i never really considered how LITTLE I knew about VR headsets and would massively explain why I get so many issues while VR sim racing. I was aware I needed to upgrade my pc but I didn't think it was as big of an issue as it really is. I have quest 2 and it's been great but It does tax my pc when racing in the rain especially and I just hadn't really fully considered that instead of getting the quest 3, I should upgrade my pc instead. sub earned!
You are one of a few people who know what you are talking about. Thank you. Let me add. My intensive experience is also frustrating. The complexity is even greater as VR performance varies per track, and the biggest problem is consistency. In my VR quest 3, some tracks are awesome, but at one turn, I get that annoying stuttering. My conclusion is that with a less perfect setup, it is kind of too hard to achieve a good VR experience. Lowering settings does not eliminate issues but just lowers them a little. You need a 4090, which I have, and the fastest CPU available with wifi 6e .... to have a near-optimal experience. VR is still a way from being perfect. Let's not forget that iRacing may need to optimize some of the tracks for VR, especially during fog and rain. My experience is the best VR projection software is Steam, but it requires a lot of CPU power. Note that Steam shines, especially with openXR. Still, the best for iRacing is the quest link cable, with OpenXR, which provides the lowest latency. Once I upgrade my CPU, I should use Quest link cable+, openXR toolkit, +Steam for projection, and iRacing, all on maxed-up settings. Yes, the experience is close to real driving 🚗. A small interesting observation is once latency is low, the control of the car becomes so amazing. That is what all top racers have, but they avoid mentioning it much. You feel immediately once the car oversteers, and correcting that becomes much easier. Sure not less than 120fps.
Thanks for the analysis. The cpu and gpu frametime comparisons are super helpful. For me not having all car mirrors is something I could live with but no dynamic shadows just totally kills the immersion for me. And if I turn on dynamic shadows I’m getting fps drops with a 4080 and 5800x3d 😢. I hope that the next gen of cpus solve this issue.
Dynamic objects ? its a killer for sure , and I suspect if used with other "expensive" settings like rain , ssao , aa at 8x performance will be horrid.
@@paulgrillo Jeah sorry that's what I meant. Ist been a while since I played around with the settings. But I think we are getting closer since I saw some TH-camr having it enabled with their 4090 and latest Intel CPUs, but that was before rain 😅 So the never ending wait for the perfect VR experience continues.
After big season updates, it can help to run a 'Graphics Config' reset. Backup your renderer.ini files (if you want). Within the GUI, select your preferred Display in the top, then go to the bottom left "Settings" --> "Interface" --> "Run Graphics Config". Then when you launch, it should reset everything.
Rather than sticking your camera in the headset, you can toggle SteamVR to display exactly what it’s outputting to your headset in a separate desktop window. Just click the hamburger menu in SteamVR to enable.
I subbed! Great vid. I have a 7800 XT myself and I can tell you it is absolutely amazing. It will blow you away! Like it has some serious FPS numbers. Would love to see you getting the card when you get 1k subs. Keep it up! Btw I got it paired with the 7800X3D CPU. Edit: added the CPU bit
Nice video! Love the explanation of how to read those charts and the simple comparison of settings. VR can be a headache sometimes getting all of the setting right, but we’ll worth the effort. I believe openXR has the same data available for free. I would love to see you explain those as well.
Thanks dude! I know how to display the overlay with OpenXR Toolkit, perhaps I can bring it up at the same time with fpsVR, and see if there's a correlation between the numbers.
Very nicely done! Now what would be awesome is to set a baseline and change each setting. This way you (and we) would know what effects the frametime in the worst way. This would set THE BASELINE for each and everyone configuring their PC for VR. See you on the track =)
Hah yeah I've thought about it, but there are some clear challenges: (1) It's unlikely that each setting will affect hardware in the same way (4080S vs 6700XT, or 13700K vs Ryzen 2700). (2) Each track + car + weather + time-of-day combination is it's own unique world, and it's unlikely the graphic settings affect performance uniformly across all of those worlds. (3) The combination of different settings probably influences the results of an individual setting. For example, say I run everything on/high and test SSR options... it's unlikely the performance impact will be the same as if I picked iRacing's Class 1 settings. It's difficult to see the forest through the trees, and impossible when you start analyzing all the species of life living within 1 square kilometer. You get lost :S
@@BenchmarkOdysseys Yea, I totally understand this. The magnitude of connections might be overwhelming. But as another TH-camr found out that changing eg. the sky does only have a small effect, MSAA up to x8 is OK, too, but MSAA x16 is a total killer. Having an Ai race with lots of Mx5 in the dry will be the starting point for everyone. so I would take this and "medium" settings as a baseline. Thinking about your research, even a different CPU/GPU combination could affect settings :D good luck in your journey, definitely got a new follower!
Hey, i am happy to come on this channel as I am also experienting a lot with Iracing with my Reverb G2 and a 7800X3d + RTX 4080. For me I have the feeling most settings dont have a big impact just some of them, but also some tracks are really different. For example VIR has terrible fps even if I put a lot of settings down and turn down my Supersampling. I also use OpenXR but native without Steam, when I used it with Steam I had the feeling i got worse performance, the only good thing was that I could use Racelabs but I found a way for that also with openxr, so just some things to consider testing on your channel. I am always working on finding the perfect setting.
@@hunterh1033 there is a program called open kneeboard from the fligh sim guys that i adjusted a little bit. its not perfect but it does the job but i can only use one overlay at a time
For iRacing rain I knocked it down to Medium shaders and got great performance, and the look of the surface changes was easier to discern. Vario Aero 4090
thanks a lot, that's a huge input with 120fps/8ms and frametimes as standard of benchmarking, as I advocate for throwing away average fps benchmarking and switching to 1% low fps for monitors/triples and 0,1% low fps for vr. As for 4 mirrors drops I guess that is one cpu core preparing data for 4 different renders(almost same as running 4 clients simultaneously) and it's not "optimized" for 4 at a time
Thanks mate! Yeah, iRacing mirrors are the 'worst' performing compared to the competition (AC, ACC, AMS2). Definitely an area that they should seek optimization.
Great work, as usual. In my case I run a 7800X3d +RTX 3080 10gb + HPG2 and openxr. I use openxr toolkit to check performance while on track. I all the scenarios that I've tested (Shadows on and off, ssr off, low or high) my system is limited by the gpu. I can have stabe 90fps by setting SSR off. CPU frametime is around 3,5 ms while gpu around 7 ms. When I turn on overlays using OpenkneeBoard GPU increases to around 8 ms. Maybe because of lack of VRAM.
So nice to see a future huge channel on the making. Amazing content. Let me ask you a question: What do you feel is the lowest-end video card capable of high graphics @ 90fps in iRacing (meta quest 3), considering equality of the rest of the hardware? And what about ACC? Do you feel geforce cards are mandatory? Thanks a lot in advance and congrats on the channel! Definitely earned a sub!
I've only completed preliminary testing with ACC, and it was discouraging! There seems to be fundamental hiccups in that sims processing of VR, whether its with GeForce or Radeon. I will try and take a closer look. Minimum hardware for VR is an interesting subject. I'm brainstorming how to compare my 3060i vs 3080ti vs 4080S in VR. Obviously the latest and greatest is 'the best', and I can show bar graphs and frametime line charts to showcase that. But a good VR experience can still be obtained on the 3060ti; it all comes down to image quality. It just requires iRacing settings to be dropped, headset resolution to be lowered, perhaps Nvidia tuning, etc. It would be great to show the side-by-side 'in the headset' video of the 3060ti to the 4080S. It will be easier to do that with a Quest 3 because of the pancake lenses. The fresnel lenses in my Valve Index really distort any camera footage I tried to capture.
@@BenchmarkOdysseys Thank you very much for such detailed response! I'm currently playing with a 3070 laptop (alienware) + Meta Quest 3. It can run Iracing somewhat smootly after a lot of tweaking on the Oculus software. ACC is painful and requires the lowest possible settings to run, but it gets so serrated and ugly i gave up. The main problem - on the laptop - is temperature. The 3070 gets hot really fast and then the stuttering and frame dropping begin. I could only solve that (on iracing) by reducing the graphics to the lowest possible option + some tweaking on the VR config file. I'm planning on moving on to a desktop build, and was thinking maybe a 4060ti with a good CPU would be enough to play in high settings in iRacing, so any benchmarking would be lovely to watch. So far i'm pretty convinced that if i want good graphics and also to hit Quest 3 120hz cap, nothing short of the 4070ti would be enough. Anyway, your CPU VR video was amazing and opened my mind to several different configurations i was not thinking before. I was really surprised to see the i5 being able to pump the 4080 to 90fps in some of your testing. Very interesting and useful knowledge. Thanks a lot!
Great video. How are you getting those graphical outputs? Particularly from Example 2? I bought fpsVR and can see their very basic history viewer but not the outputs you show on your video.
No prob, glad to hear you're trying out the tool :) The fpsVR History Viewer is a separate application found within the installation folder for fpsVR. It will present the same charts as I've shown.
Great videos! Could we get a video comparing the gpus from 20 series to 40 series using the same cpu? for exemple i9 13900k or 14900 or ryzen 9 7950x3d and 7800x3d?
Thanks dude. I think the most efficient software for delivering VR to my Valve Index is with the SteamVR. I tried CapFrameX with the Oculus, but as you know, the results shown are not very representative of what's actually occurring because the GPU streams the feed to the HMD.
This is amazing I learned so much thank you. Just started with VR and I'm surprised you dont get any late frames at 90hz with 2 mirrors and pretty high settings. I'm having a hard time not to have any stutters at 90 while not having a single mirror enabled (7800x3d 4080s). Have quest 3 tho but higher res would mostly impact gpu no? My gpu is not having any issue.
Actually, one thing I was considering is the fact that an HMD on our head usually leads to increased draw calls, particularly because of the constant head movement and, because sim racing, the change in object draw. *example: looking left/right and then back while also driving forward really fast* Would that have any impact on benchmarking? Or am I just misunderstanding the concept of draw calls here?
That was one of my original concerns. But after several tests, I could not measure a difference between static position, normal head rotation while racing, and exaggerated bobble-head movements. It also doesn't matter if the mirrors can be visually seen on the HMD or not. I don't think the iRacing graphics engine is sophisticated enough do reduce rendering workload of a mirror when it's out of sight.
We got rain, everyone wanted it, but a season later, no one wants to race in it... Falken Sports Car Challenge has 100% rain this week and its been one split almost every race...
Last week I watched the 6 hours at Watkins. When the rain comes down hard, instantly 3 or 4 cars go off with hydroplaning, and it's full course yellow for +20mins. Which gives all the teams the opportunity to pit for wet tires and adapt to the conditions under safety car speed, which also helps to develop a dry line. I think iRacing has captured what it's like to drive in the rain. Even with a gradual transition to rain, or configuring steady conditions for a session, it is frustratingly difficult. With the driver interviews from the Glen, some drivers loved the thrill and challenge, others couldn't wait to complete their stint and get out of the car ;)
@@BenchmarkOdysseys I decided to do the Falken challenge last night, one split and was getting like 40fps on my quest 2 (7800x3d/4080), I wanted it to end so badly but I finished and even made it with only one incident point. Directly after that I went and did the GT4 fixed on the same track and was absolutely flying. Driving in the rain makes you a better driver, I’m convinced of that now
You're welcome :) I've got a good understanding of how to tune iRacing and read the data results, so I'll do the first GPU comparison with VR using it. I'd like to ACC next, however, that sim seems to have fundamental issues with processing VR, even with high-end hardware like 7800X3D and 4080S.
Valve ditched OpenVR a few years ago, so SteamVR natively uses OpenXR API. The 'FPS' presented in fspVR seems to directly match the in-sim 'FPS' performance that iRacing reports. From my understanding, iRacing's R and G meters are internal metrics just for the sim's performance; they do not report on the bigger picture of what is happening to reproduce 3D rendering within the VR HMD.
Please excuse the stupid question, but do we want combined CPU + GPU frametimes beneath refresh? Or just both individually? E.g. I have a 80 Hz headset and need to be beneath 12.5ms. Can I run CPU at 10 ms and GPU at 10 ms (assuming neither have tails beyond 12.5ms)? Or do I have to keep the sum of CPU and GPU beneath 12.5ms? At say 4ms CPU and 8ms GPU?
Within fpsVR, the CPU and GPU frametimes are presented individually. Ideally CPU is less than 12.5ms at all times and GPU is less than 12.5ms at all times, if rendering 80Hz. If either component is overwhelmed, it can drag the other with it. Note that these values are different than what is reported within iRacing's R and G meters. Those can actually be misleading.
@@BenchmarkOdysseys Brilliant, thanks for clarifying! I'm guessing there's always a frame latency then with the VR vsync which allows the CPU to work on the next frame while the GPU works on the previous frame. I've been using the CPU and GPU times given by the OpenXRToolkit overlay. Otherwise I think the R meter in iRacing is pegged at 12.5ms when using OpenXR.
Technically some ROG Ally models meet iRacing minimum specs (4 cores, 16gb ram, 4gb vram, 25gb storage). You just need to figure out how to strap it to your face B)
Hey dudes. I’m interested in getting involved in Iracing (PS5-Gran Turismo PSVR2 guy here), and I’m currently spec-ing out Pre-built PC’s. I think a quest 3 will suit my VR needs. Is it still advantageous, or necessary to build out a high-spec PC with a 7800 3DX or 13th gen intel CPU, along with a 4090 graphics card? I don’t overspending per se, if it buys me a bit of “future-proofing”. But I don’t want you to needlessly throw money away either. I honestly have no clue what I’m looking at with all these letters and numbers on PC spec sheets 😂
Apparently PSVR2 support on Windows is being developed by Sony, and they expect to release the drivers and/or software this year. Not sure if it will be worth waiting for. But if you'd like to talk system builds, flip me an email at benchmarkodysseys@gmail.com :)
I run my system with meta 3. 4070ti oc 2970mhz and 5800x 4900mhz. About 90 fps, in rain 70-90(just make some on low) I didnt like fsr or dlss in vr. And dont use steam vr. Openxr with cable 700mbits is a killer
Hmm, my english is not that good...i own a 7800x3d / RTX4090 PC and even when running iRacing having turned everything (!) off, i get stutters...when i watch (from the pitlane camera) cars passing by, they stutter. Why?
Hello! Sorry to hear about your stutters, I wish iRacing and VR was easier to configure. There many (potential) mistakes you can make when configuring your hardware and selecting iRacing options. If you haven't already, I recommend you go to the official iRacing forums and start a thread asking for help. You will quickly find help there by sharing as much as you can (all your system specs, iRacing config, VR software config, etc).
I have a 4070ti super and 14900kf and iracing freezes a lot when using vr. The P bar spikes and the screen will freeze for like 2-5 seconds. The R G and T are all running pretty high as well. They’re grey, but the bars are all very large. When the screen freezes, those will spike and all turn red. Then it settles down and I’m fine again for a while. So it just randomly spikes like crazy. I have no clue what to do to fix this. Any help would be appreciated. My fpsvr graphs show both GPU and CPU are barely being taxed…… no clue what the deal is
![BOWKYLION - วิงวอน (ex-change) [Official MV]](http://i.ytimg.com/vi/cLdnZWDaO-w/mqdefault.jpg)
My man, you don’t have a clue how often I asked several of your fellow TH-camrs to talk about VR performance when looking and benching CPU’s and GPU’s. Who cares anymore if a system can poop out 450FPS or 400 FPS on a flat screen. This is soooo appreciated and the way you have approached this is second to none. This is the most clear explained and visualized bench I have ever seen and you got my sub for that. This is great man and I foresee greatness in your channel. Keep it up. Kuddo’s
Thanks, this clears up a lot for me. I could never really grasp how it works. I subbed, I love when people explain the technical side of iRacing. I'm a mechinical engineer myself, so I understand the car side of the sim but the computer/software side is some kind of black magic. Keep going, you deserve those 1000 subs.
Thanks for explaining the new presentation of the benchmarks. Quality information as always.
Here's to 1K subs! 🍻
Fantastic explanation and I appreciate you for laying out the thought process. Thank you!
Really liking the VR benchmarks with SIm Racing, Keep it up!
Clear explanation 👍 The examples at the end were very interesting. I'll definitely be re-thinking having cockpit mirrors. And exploring ASW. Have never turned it on, but it looks to do a very good job of reducing frametimes below the 12ish ms I need for my 80 Hz Rift S.
That is a mega effort on the benchmarks , explanation , and presentation. Hopefully you get to 1000 subs easy , i added 1 :)
Thank you for this video. Finally I understand why I'm seeing stutters in VR, I couldn't understand why I had a solid FPS but it didn't feel smooth. Your stutter example at the start of the video is exactly what I am seeing. Thank you also for explaining in simple terms how to interpret the results of your testing. I've got a new 7800X3D arriving today, so hopefully that will help reduce my stutters.
Great info and perfectly timed for me. Thank you. Looking forward to the next installment.
This is great. I'm quite competent with computers most of the time unless it's something really complicated like actually coding or whatever and I am NOT a professional by any means, I just use pcs a lot and am usually the one to host servers, create mod packs for games for my friends etc but i never really considered how LITTLE I knew about VR headsets and would massively explain why I get so many issues while VR sim racing. I was aware I needed to upgrade my pc but I didn't think it was as big of an issue as it really is. I have quest 2 and it's been great but It does tax my pc when racing in the rain especially and I just hadn't really fully considered that instead of getting the quest 3, I should upgrade my pc instead. sub earned!
You are one of a few people who know what you are talking about. Thank you.
Let me add. My intensive experience is also frustrating. The complexity is even greater as VR performance varies per track, and the biggest problem is consistency. In my VR quest 3, some tracks are awesome, but at one turn, I get that annoying stuttering. My conclusion is that with a less perfect setup, it is kind of too hard to achieve a good VR experience. Lowering settings does not eliminate issues but just lowers them a little.
You need a 4090, which I have, and the fastest CPU available with wifi 6e .... to have a near-optimal experience. VR is still a way from being perfect. Let's not forget that iRacing may need to optimize some of the tracks for VR, especially during fog and rain.
My experience is the best VR projection software is Steam, but it requires a lot of CPU power. Note that Steam shines, especially with openXR. Still, the best for iRacing is the quest link cable, with OpenXR, which provides the lowest latency. Once I upgrade my CPU, I should use Quest link cable+, openXR toolkit, +Steam for projection, and iRacing, all on maxed-up settings. Yes, the experience is close to real driving 🚗. A small interesting observation is once latency is low, the control of the car becomes so amazing. That is what all top racers have, but they avoid mentioning it much. You feel immediately once the car oversteers, and correcting that becomes much easier. Sure not less than 120fps.
By the way, there is a setting in fpsvr that you can enable the desktop window which will also show the exact same thing
Outstanding as always - really well presented complex data, your like the digital foundry of sim racing :-)
Turning off motion smoothing was all I needed from ya cause everyone misses that. But so far sounds good. Subbed. :)
How do you turn off motion smoothing? Can’t see the options
There is also a built in performance overlay feature in virtual desktop for quest users. No need to spend extra money on a separate app.
well made content, im mostly not commenting anything but dude those charts and everything is super clear, keep it up! :)
Thanks for the analysis. The cpu and gpu frametime comparisons are super helpful.
For me not having all car mirrors is something I could live with but no dynamic shadows just totally kills the immersion for me. And if I turn on dynamic shadows I’m getting fps drops with a 4080 and 5800x3d 😢. I hope that the next gen of cpus solve this issue.
Dynamic objects ? its a killer for sure , and I suspect if used with other "expensive" settings like rain , ssao , aa at 8x performance will be horrid.
@@paulgrillo Jeah sorry that's what I meant. Ist been a while since I played around with the settings.
But I think we are getting closer since I saw some TH-camr having it enabled with their 4090 and latest Intel CPUs, but that was before rain 😅
So the never ending wait for the perfect VR experience continues.
After big season updates, it can help to run a 'Graphics Config' reset. Backup your renderer.ini files (if you want). Within the GUI, select your preferred Display in the top, then go to the bottom left "Settings" --> "Interface" --> "Run Graphics Config". Then when you launch, it should reset everything.
Iv got same setup
This channel is so cool! thanks for your contribution
I enjoyed the detail in this, I haven't used SteamVR for ages and did not know it would run the OpenXR API, nice to know, may give it a shot, cheers.
Rather than sticking your camera in the headset, you can toggle SteamVR to display exactly what it’s outputting to your headset in a separate desktop window. Just click the hamburger menu in SteamVR to enable.
I subbed! Great vid. I have a 7800 XT myself and I can tell you it is absolutely amazing. It will blow you away! Like it has some serious FPS numbers. Would love to see you getting the card when you get 1k subs. Keep it up!
Btw I got it paired with the 7800X3D CPU.
Edit: added the CPU bit
Nice video! Love the explanation of how to read those charts and the simple comparison of settings. VR can be a headache sometimes getting all of the setting right, but we’ll worth the effort.
I believe openXR has the same data available for free. I would love to see you explain those as well.
Thanks dude! I know how to display the overlay with OpenXR Toolkit, perhaps I can bring it up at the same time with fpsVR, and see if there's a correlation between the numbers.
Very nicely done!
Now what would be awesome is to set a baseline and change each setting.
This way you (and we) would know what effects the frametime in the worst way.
This would set THE BASELINE for each and everyone configuring their PC for VR.
See you on the track =)
Hah yeah I've thought about it, but there are some clear challenges:
(1) It's unlikely that each setting will affect hardware in the same way (4080S vs 6700XT, or 13700K vs Ryzen 2700).
(2) Each track + car + weather + time-of-day combination is it's own unique world, and it's unlikely the graphic settings affect performance uniformly across all of those worlds.
(3) The combination of different settings probably influences the results of an individual setting. For example, say I run everything on/high and test SSR options... it's unlikely the performance impact will be the same as if I picked iRacing's Class 1 settings.
It's difficult to see the forest through the trees, and impossible when you start analyzing all the species of life living within 1 square kilometer. You get lost :S
@@BenchmarkOdysseys Yea, I totally understand this. The magnitude of connections might be overwhelming. But as another TH-camr found out that changing eg. the sky does only have a small effect, MSAA up to x8 is OK, too, but MSAA x16 is a total killer.
Having an Ai race with lots of Mx5 in the dry will be the starting point for everyone. so I would take this and "medium" settings as a baseline.
Thinking about your research, even a different CPU/GPU combination could affect settings :D
good luck in your journey, definitely got a new follower!
what a great video! thank you for all the excellent info!
Information of very good quality, thanks a lot !!!
Hey, i am happy to come on this channel as I am also experienting a lot with Iracing with my Reverb G2 and a 7800X3d + RTX 4080. For me I have the feeling most settings dont have a big impact just some of them, but also some tracks are really different. For example VIR has terrible fps even if I put a lot of settings down and turn down my Supersampling. I also use OpenXR but native without Steam, when I used it with Steam I had the feeling i got worse performance, the only good thing was that I could use Racelabs but I found a way for that also with openxr, so just some things to consider testing on your channel. I am always working on finding the perfect setting.
How did you manage racelabs while using OpenXR native if you don't mind sharing?
@@hunterh1033 there is a program called open kneeboard from the fligh sim guys that i adjusted a little bit. its not perfect but it does the job but i can only use one overlay at a time
Look for OpenKneeBoard
I wish there were some ACC VR stuff, still cool to see everything of course.
Thanks, great video! Super informative as always
I subbed for the 7900xtx VR performance video!
Amazing content, subbed!
For iRacing rain I knocked it down to Medium shaders and got great performance, and the look of the surface changes was easier to discern. Vario Aero 4090
thanks a lot, that's a huge input with 120fps/8ms and frametimes as standard of benchmarking, as I advocate for throwing away average fps benchmarking and switching to 1% low fps for monitors/triples and 0,1% low fps for vr.
As for 4 mirrors drops I guess that is one cpu core preparing data for 4 different renders(almost same as running 4 clients simultaneously) and it's not "optimized" for 4 at a time
Thanks mate! Yeah, iRacing mirrors are the 'worst' performing compared to the competition (AC, ACC, AMS2). Definitely an area that they should seek optimization.
Great work, as usual. In my case I run a 7800X3d +RTX 3080 10gb + HPG2 and openxr. I use openxr toolkit to check performance while on track. I all the scenarios that I've tested (Shadows on and off, ssr off, low or high) my system is limited by the gpu. I can have stabe 90fps by setting SSR off. CPU frametime is around 3,5 ms while gpu around 7 ms. When I turn on overlays using OpenkneeBoard GPU increases to around 8 ms. Maybe because of lack of VRAM.
Outstanding tutorial mate, thank you
So nice to see a future huge channel on the making. Amazing content.
Let me ask you a question: What do you feel is the lowest-end video card capable of high graphics @ 90fps in iRacing (meta quest 3), considering equality of the rest of the hardware? And what about ACC? Do you feel geforce cards are mandatory?
Thanks a lot in advance and congrats on the channel! Definitely earned a sub!
I've only completed preliminary testing with ACC, and it was discouraging! There seems to be fundamental hiccups in that sims processing of VR, whether its with GeForce or Radeon. I will try and take a closer look.
Minimum hardware for VR is an interesting subject. I'm brainstorming how to compare my 3060i vs 3080ti vs 4080S in VR. Obviously the latest and greatest is 'the best', and I can show bar graphs and frametime line charts to showcase that.
But a good VR experience can still be obtained on the 3060ti; it all comes down to image quality. It just requires iRacing settings to be dropped, headset resolution to be lowered, perhaps Nvidia tuning, etc. It would be great to show the side-by-side 'in the headset' video of the 3060ti to the 4080S. It will be easier to do that with a Quest 3 because of the pancake lenses. The fresnel lenses in my Valve Index really distort any camera footage I tried to capture.
@@BenchmarkOdysseys Thank you very much for such detailed response!
I'm currently playing with a 3070 laptop (alienware) + Meta Quest 3. It can run Iracing somewhat smootly after a lot of tweaking on the Oculus software. ACC is painful and requires the lowest possible settings to run, but it gets so serrated and ugly i gave up.
The main problem - on the laptop - is temperature. The 3070 gets hot really fast and then the stuttering and frame dropping begin. I could only solve that (on iracing) by reducing the graphics to the lowest possible option + some tweaking on the VR config file.
I'm planning on moving on to a desktop build, and was thinking maybe a 4060ti with a good CPU would be enough to play in high settings in iRacing, so any benchmarking would be lovely to watch.
So far i'm pretty convinced that if i want good graphics and also to hit Quest 3 120hz cap, nothing short of the 4070ti would be enough.
Anyway, your CPU VR video was amazing and opened my mind to several different configurations i was not thinking before. I was really surprised to see the i5 being able to pump the 4080 to 90fps in some of your testing. Very interesting and useful knowledge.
Thanks a lot!
Great video. How are you getting those graphical outputs? Particularly from Example 2? I bought fpsVR and can see their very basic history viewer but not the outputs you show on your video.
No prob, glad to hear you're trying out the tool :) The fpsVR History Viewer is a separate application found within the installation folder for fpsVR. It will present the same charts as I've shown.
Great videos! Could we get a video comparing the gpus from 20 series to 40 series using the same cpu? for exemple i9 13900k or 14900 or ryzen 9 7950x3d and 7800x3d?
Great Video!
You could try to do the bench’s with CapFrameX & OpenXR without SteamVR. 🙌🏻
Thanks dude. I think the most efficient software for delivering VR to my Valve Index is with the SteamVR. I tried CapFrameX with the Oculus, but as you know, the results shown are not very representative of what's actually occurring because the GPU streams the feed to the HMD.
This is amazing I learned so much thank you. Just started with VR and I'm surprised you dont get any late frames at 90hz with 2 mirrors and pretty high settings. I'm having a hard time not to have any stutters at 90 while not having a single mirror enabled (7800x3d 4080s). Have quest 3 tho but higher res would mostly impact gpu no? My gpu is not having any issue.
Actually, one thing I was considering is the fact that an HMD on our head usually leads to increased draw calls, particularly because of the constant head movement and, because sim racing, the change in object draw. *example: looking left/right and then back while also driving forward really fast*
Would that have any impact on benchmarking? Or am I just misunderstanding the concept of draw calls here?
That was one of my original concerns. But after several tests, I could not measure a difference between static position, normal head rotation while racing, and exaggerated bobble-head movements.
It also doesn't matter if the mirrors can be visually seen on the HMD or not. I don't think the iRacing graphics engine is sophisticated enough do reduce rendering workload of a mirror when it's out of sight.
Great content l, thank you! Unfortunately I can’t sub again but I am sure you will get to 1k subs in no time.
Was wondering if sidequest for oculus could decrease headset latency?
We got rain, everyone wanted it, but a season later, no one wants to race in it... Falken Sports Car Challenge has 100% rain this week and its been one split almost every race...
Last week I watched the 6 hours at Watkins. When the rain comes down hard, instantly 3 or 4 cars go off with hydroplaning, and it's full course yellow for +20mins. Which gives all the teams the opportunity to pit for wet tires and adapt to the conditions under safety car speed, which also helps to develop a dry line.
I think iRacing has captured what it's like to drive in the rain. Even with a gradual transition to rain, or configuring steady conditions for a session, it is frustratingly difficult.
With the driver interviews from the Glen, some drivers loved the thrill and challenge, others couldn't wait to complete their stint and get out of the car ;)
@@BenchmarkOdysseys I decided to do the Falken challenge last night, one split and was getting like 40fps on my quest 2 (7800x3d/4080), I wanted it to end so badly but I finished and even made it with only one incident point. Directly after that I went and did the GT4 fixed on the same track and was absolutely flying. Driving in the rain makes you a better driver, I’m convinced of that now
And about do this test on AC or ACC ? thanks for the video
You're welcome :) I've got a good understanding of how to tune iRacing and read the data results, so I'll do the first GPU comparison with VR using it. I'd like to ACC next, however, that sim seems to have fundamental issues with processing VR, even with high-end hardware like 7800X3D and 4080S.
@@BenchmarkOdysseys I’m about to get a 5600x and rtx 3060ti with a pico 4 or reverb g2, for iracing so you think this is enough?
Great video
Hi. Thank you for the video. How do you use fpsvr with openxr?
Valve ditched OpenVR a few years ago, so SteamVR natively uses OpenXR API. The 'FPS' presented in fspVR seems to directly match the in-sim 'FPS' performance that iRacing reports. From my understanding, iRacing's R and G meters are internal metrics just for the sim's performance; they do not report on the bigger picture of what is happening to reproduce 3D rendering within the VR HMD.
Great content ❤
Please excuse the stupid question, but do we want combined CPU + GPU frametimes beneath refresh? Or just both individually? E.g. I have a 80 Hz headset and need to be beneath 12.5ms. Can I run CPU at 10 ms and GPU at 10 ms (assuming neither have tails beyond 12.5ms)? Or do I have to keep the sum of CPU and GPU beneath 12.5ms? At say 4ms CPU and 8ms GPU?
Within fpsVR, the CPU and GPU frametimes are presented individually. Ideally CPU is less than 12.5ms at all times and GPU is less than 12.5ms at all times, if rendering 80Hz. If either component is overwhelmed, it can drag the other with it.
Note that these values are different than what is reported within iRacing's R and G meters. Those can actually be misleading.
@@BenchmarkOdysseys Brilliant, thanks for clarifying! I'm guessing there's always a frame latency then with the VR vsync which allows the CPU to work on the next frame while the GPU works on the previous frame.
I've been using the CPU and GPU times given by the OpenXRToolkit overlay. Otherwise I think the R meter in iRacing is pegged at 12.5ms when using OpenXR.
I subbed!!
thanks dude!
Guys, I have a doubt. Can I play Iracing in vr with my rog ally??
Technically some ROG Ally models meet iRacing minimum specs (4 cores, 16gb ram, 4gb vram, 25gb storage). You just need to figure out how to strap it to your face B)
So if I have a 4070 super and an i7 14700 and 32 gb Ram I should be good?
That would be a capable system!
Hey dudes. I’m interested in getting involved in Iracing (PS5-Gran Turismo PSVR2 guy here), and I’m currently spec-ing out Pre-built PC’s.
I think a quest 3 will suit my VR needs. Is it still advantageous, or necessary to build out a high-spec PC with a 7800 3DX or 13th gen intel CPU, along with a 4090 graphics card? I don’t overspending per se, if it buys me a bit of “future-proofing”. But I don’t want you to needlessly throw money away either.
I honestly have no clue what I’m looking at with all these letters and numbers on PC spec sheets 😂
Apparently PSVR2 support on Windows is being developed by Sony, and they expect to release the drivers and/or software this year. Not sure if it will be worth waiting for. But if you'd like to talk system builds, flip me an email at benchmarkodysseys@gmail.com :)
I run my system with meta 3. 4070ti oc 2970mhz and 5800x 4900mhz. About 90 fps, in rain 70-90(just make some on low)
I didnt like fsr or dlss in vr.
And dont use steam vr. Openxr with cable 700mbits is a killer
Adaptive sinc is off(need every pc restart)
And 1 or 0 mirrors😂
Hmm, my english is not that good...i own a 7800x3d / RTX4090 PC and even when running iRacing having turned everything (!) off, i get stutters...when i watch (from the pitlane camera) cars passing by, they stutter. Why?
Hello! Sorry to hear about your stutters, I wish iRacing and VR was easier to configure. There many (potential) mistakes you can make when configuring your hardware and selecting iRacing options. If you haven't already, I recommend you go to the official iRacing forums and start a thread asking for help. You will quickly find help there by sharing as much as you can (all your system specs, iRacing config, VR software config, etc).
I have a 4070ti super and 14900kf and iracing freezes a lot when using vr. The P bar spikes and the screen will freeze for like 2-5 seconds. The R G and T are all running pretty high as well. They’re grey, but the bars are all very large. When the screen freezes, those will spike and all turn red. Then it settles down and I’m fine again for a while. So it just randomly spikes like crazy. I have no clue what to do to fix this. Any help would be appreciated. My fpsvr graphs show both GPU and CPU are barely being taxed…… no clue what the deal is