Hippo

Thank you. Sorry for wasting everyone's time and for my disturbing lack of faith; if it's any consolation, I'm feeling rather stoopid now for not having tried that. When I turn the brightness down I can clearly(ahem) see it is indeed transparent. I will turn down the brightness in future when required. EDIT: Sorry to be a pain, but having thought about it a bit more, I still feel it's not quite right. The transparency goes down and the shade of green gets "brighter" as the HUD brightness is increased, so that it is virtually opaque at the default brightness setting. Would having the transparency not go down so much while the shade of green gets "brighter" be a possibility? I do realise that this suggestion might, technically, be nonsense.

I'm using VR. Sorry if I've got this wrong but I've been wondering about it for ages. I'm not sure this always happens but it seems to me that the symbology on the hud is opaque and it shouldn't be. E.g. when I place the ASL over a target at distance (i.e. a dot) to try and line up my approach, I can't see the target "through" the ASL which means I lose sight of the target.

Great post, thank you.

Thanks. Is the structure/ available variables/ etc of mission files documented anywhere, or have people, such as those who have developed mission generators, just had to work it all out for themselves?

E.g. I have 10 mission files and have made multiple individual image files for each and I wish to add them as briefing images? Is there any way to do this other than manually using the ME? Or change the weather? Or add triggers? Etc.

I very recently upgraded from a 1080ti to a 3080ti. I am still tuning, but currently I am running with these settings. My aim is to run at 72 fps w/o ASW kicking in on moderately simple SP missions (was not possible with 1080ti). I would like to run at 1.3 - 1.5 PD, but I find that I can't hold 72 fps, as I'd like, if I do, so am running at 1.1. I think it appers considerably crisper, esp for small text, in MFDs, etc, and would love to run it this high, but it impacts my performance too much. I am still running with my i7-8700k, and will be upgrading this to something newer soon (prob 13700k), which I hope will squeeze out a bit more performance. I havent tried going higher that 72 Hz, which again would be nice, by will of course impact performance. I'm actually very happy with the upgrade, as I highly value being able to run w/o ASW.

https://developer.oculus.com/blog/asynchronous-spacewarp/?locale=en_GB

Thanks again, many great points there. At the risk of being tedious, did you try my track with the ST exe, and was performance destroyed or noticeably worse? I think this is going over my head. What is meant by CPU latency? Am I understanding correctly that the likelyhood of CPU bottlenecks tanking frame rates in specific situations with very high settings is greatly reduced in MT, but that in situations like my track with lowish settings like mine it will make no noticeable difference to frame rates? I refer you again to the YT video I linked to. That guy has a top end system meeting all of your requirements above and yet sees no difference between the ST and MT versions of the exe. But I take your point on focusing exclusively on the fps number. The ED FAQ states: I can find numerous pancake YT videos demonstrating fps increases varying from 50%-100%, yet am unable to find VR videos showing similar gains. My own view is that the above statement is at worst incorrect, and at best needs to be qualified (as to what exactly they mean by "performance").

Done

Quest 2, 8700k, 3080ti. In the MT version, when over the island (towards the end of attached track) and in a bank when I look at the terrain going past me (i.e. looking to my left or right) I see very visible terrrain stuttering which I don't see at all in the ST version. Frame rates (according to the ingame fps counter) are stable at either 72 fps native or 36 asw. I experienced this while flying the mission or replaying the track. Look here for futher detail. mt_test_20230311.trk

Firstly: thank you so much for taking the time to do that. I am unfamiliar with the G2 so I assume that it runs "native" at 90 Hz, and drops to 45 Hz with reproj when the system can't keep up. Reprojection is acceptable for me, up to a point. I can live with it in helos, but the artifacting it introduces in fast jets as maneuvers can be much more violent is a problem. Ideally I would want not reproj, or for it only to kick under very heavy loads. I prefer to drop the eye candy to maintain native if possible. I understand this point is very subjective. Maybe the G2 does reproj better? And throwing FSR into the mix? For me, I was able to maintain 72fps (native) throughtout (i.e. I could not cause a single drop into reproj) with MT. With ST I could induce ocassional drops to reproj if I looked around while flying close to the ground. Unfortunately in MT, I could see visible terrain stuttering when looking out to my left/ right while banking over the island, even though I was running at a constant 72 Hz. I did not see this stuttering with ST. Sorry to be pedantic, but could you unequivocally confirm that you did not experience this? Forgive me, but I am unsure as to what exactly your frame times are telling you. Unless I am misunderstanding things frame time is just the inverse of fps, and has the advantage that comparisons between numbers are in direct proportion. I am old school (or just old) and have been using fps for decades, so have a good "feel" for what they represent in reality, and continue to use fps in comparisons, though I probably should switch. As another poster very reasonably suggests, if the cpu frame time to which you refer is a theoretical max whilst in reality what you are actually seeing is limited by your gpu frametime then I would suggest that it is not all that relevant. I hate to impose on your good nature, but what would be really useful is if you could also run the track ST and MT and see if you notice any differences in fps, especially when looking directly at the ground beneath you when the plane is in a bank. For extra credit it you could fire up msi afterburner and screenshot a graph of what all your cores and gpu %util are doing throughout, this would be very useful information to diagnose why you are seeing fps improvements that others are not. It is only where the aircraft is over the island that the system is under stress, so a pretty poor choice of track from me... sorry. This chap demonstrates a performance increase between an older ST version and the current MT version, but what is interesting is that he states in the comments that he's noticed no difference between the current ST and MT versions in VR. The concerning implication is that he would have seen exactly the same increase if he had compared the new ST version with the old ST version, and one has to be wonder whether any improvements people are noticing in VR are down to other changes and not MT itself. I unfortunately did not run tests with the version before the update, but from memory I'm almost certain that performance was considerably worse. Most (all?) of the other YT videos that I've seen demonstrating large performance increases are in pancake. I don't doubt that you are seeing what you say, and again I'm sorry if any of my remarks came across as snarky - it really isn't my intention, but from all the other evidence I've seen, VR performance improvements appear to be marginal at best. Edit and note: I notice that many posters say things like I got 35 fps before and now I get 50 fps. I can only assume that they are running w/o reprojection, i.e. with unlocked constantly varying fps. I gave up on doing this years ago as I found it too uncomfortable, and always run with reproj, so I only ever see 36/72 fps, which is why I compare cpu / gpu loads at the same fps.

Sorry if you've already mentioned it, I couldn't see it mentioned anywhere, but what VR hardware are you using? If you have some time to kill, I posted a track in the thread I linked to, and I'd be very interested to hear about your experience replaying that track. In particular towards the end where I'm flying low over the terrain. Look out to your side at the ground when I'm flying / banking low over it. Is it smooth without stutter? What sort of fps do you get? Is the frame rate stable? Is there any reprojection going on? Edit: the maps you mention I haven't tried. I don't own Syria, and gave up on Marianas the first time I tried it (although I had a 1080ti at the time - unusable slide show). I've only tested on Caucasus, Persian Gulf, and a bare Supercarrier on missions with very little happening. Perhaps I might see greater improvements with the maps you mention, and with missions with more going on. I did notice that the CPU was doing a lot less on the SC missions, perhaps because there was no terrain nearby.

I think that there is an improvement, as I have documented/ droned on about here. Unfortunately, it seems pretty minor, especially when compared with numerous YT videos out there for pancake with fps improvements ranging from 50%-100%. Although my initial excitement has been somewhat tempered, hopefully we will see further improvements in the future. I don't mean to offend anyone, but I've learnt to take comments such as "Multithreading Performance is Sensational in VR" with a generous pinch of salt unless they're backed up by screenshots / videos with numbers, charts, etc.

Good to hear. If you'd like to test play my track and whilst low over the island look to your left / right so that you're looking at the ground as it goes by. If I do this with MT I see stuttering (it's smooth when I'm looking forward), but it's smooth with the ST version. Maybe just me...

Time to start qualifying my original post. I did some further testing using the F/A-18 mission "Weapons Qualification - AGM65E Maverick Practice" (Persian Gulf) as flown in the attached track, with the following settings. It's when I get close to the island that the GPU starts to work. With MT i am able to fly low over the island, looking around, at 72 fps without ever going into reprojection. With ST I sometimes drop into reproj. However, when banking over the island and looking down at the ground it's smooth in ST but visibly stutters in MT (I believe this is a known issue). So, although there is a measurable performance improvement, for me it is slight and in practical usage does not make a significant difference; and there are currently drawbacks with MT. However, this is just the start and hopefully we will see further improvements going forward. i8700k, 32GB, 3080Ti, Quest 2 - I'd be very grateful if anyone with a similar GPU but much newer CPU or top of the range (e.g. 4090/13900) could fly the same mission in the same way play the track and let me know how much further they can bring up their DCS settings (at the same quest2 settings below) without reproj kicking in. I can get reproj to kick in (ST only) while over the island if I look around. mt_test_20230311.trk

Well I haven't compared shadow quality between the MT and non-MT versions, but I am measuring better performance on the MT version with the same settings. About to post an update though...

Ah, ok. Hadn't noticed, but then wasn't really paying attention.

HI - MED - LOW, other settings as above (with MT version):

I know what you mean. Although this is the first time I've been hopeful for a long time. Going with my comparisons above something has almost certainly changed in the way the CPU load is being distributed. It seems as if the cores are all doing more than before (except core 12 which previously seemed to be doing everything) and the GPU %util is down by around 10%. The good thing is that this is just the start and hopefully things can be improved. As far as I can tell shadows are working correctly, although I'm an in the cockpit almost all the time guy so external views are something I don't usually bother with; in my first test I was running with low, in the second with high, in both tests shadows were working, the ext screenshot is with shadows high.

Yes I am. I have not noticed any degradation in image quality.

I've not been keeping up with all of this openxr stuff, so I don't know which I'm using. I'm using non Steam DCS, I just run the Oculus software first, then start DCS usually from the within the Dash desktop. I'm not running any flags and haven't changed anything. Aside from an improvement in performance, things seem more stable and less "spiky", for instance, looking around in ST could often cause temporary drops into reprojection, my (very limited) experience so far with MT has been very solid. see prev post

Ran another test and raised some settings so as below. Raised the following: pd 1.0 > 1.2 visib range med > high heat blur low > high shadows low > high sec shadows off > on msaa 2x > 4x I was able to maintain 72 fps in MT all the way to touchdown, but not with ST.

As a long-time somewhat frustrated VR user, I had to pop in and say thanks ED for the update and all your efforts. My first impressions are very promising. I've only been experimenting for about half an hour, but here's a before / after landing the f-18. Core 12 has always been where I would hit my limit. Notice that Core 1 is now at 100%, whereas it was doing very little before. This was on a v simple (just my a/c) mission with lowish settings as my aim is to achieve 72 fps on Q2 w/o reprojection with spare headroom for busier scenerios. It seems as if I have more headroom now. Further experimentation needed to see how high I can get my settings. I only recently upgraded from a 1080ti to a 3080ti, and didn't get quite as big a performance bump I was hoping for. My 8700 is due for an upgrade; opinions on what sort of bump I can get going to a 13700/13900? Anyway, now back to the sim