St4rgun

As I had to switch OFF the hyperthreading (which was normally OFF for me always, but because of the MT FAQ I switched it on on recommendation - it was a bad move) now I'm back to OpenXR 300% and FSR 60% in OpenXR toolkit which works flawlessly. With FSR I suppose DCS "thinks" that my G2 has about 3200x3200 resolution per eye, that's the rendering target. Then FSR scales this up to the much higher resolution for display, but that uses only the GPU (at least I hope so).

Dear @BIGNEWY, it turned out that maybe all my bad experience related to the mindless following of the MT FAQ which says "Currently, users with more than 32 threads should disable HT/SMT, but for everyone else we recommend enabling HT/SMT." For me (10700k) enabling hyperthreading caused massive stutter and worse performance than the ST version, while switching off hyperthreaing solved these kind of problems. Maybe you should correct this FAQ item because it's not true for VR at the moment! Thanks.

I can confirm, for me disabling hyperthreading helped a lot. The extreme stuttering is gone. Now I can use again the OpenXR 300% resolution with OpenXR Toolkit"s FSR 60% for maximum clarity with my G2.

+1: green GPU bound 2D, red CPU bound main thread VR (G2).+4-5 fps overall, but much worse stutter. Motion reprojection always off, full repair.

You would be right as you saw my signature. But with the MT beta I had to turn off the 300% upscale because it was terrible (I came to the conclusion that FSR simply does not work well), so the results I wrote were already based on "only" 120% OpenXR settings without FSR. But let me repeat: 60-75 fps in the menu screen is ridicuolus. Something is really seriously wrong with this version, because the ST version menu is locked to 90 fps.

I also did a full repair before my test. Nominately the fps numbers are a bit better in MT than the ST version (5-10%), but the stuttering is much worse. It's even horrendous when looking sideways. I only tried quick take offs and free flights with the built in instant action missions on Caucasus, Nevada, Marianas maps. Interestingly the F-5E Caucasus winter take off mission with the snow became even much worse az the buildings of newar Batumi airport despite the fps counter says 25-30 fps but it is so stuttered that it's a max of 3-4 fps in reality. If I only look straight then it's much better. Even the menu produces 60-75 fps maximum (!!! with a 3090 NO WAY), while in ST version it's locked at 90 fps. Something is really bad with this build for VR...

Experience in VR / 2D with the newest MT (SP, F-16 take off Nevada instant action staying on the runway and just looking around): 42 fps CPU bound (main thread) !!! - what does that mean? I thought that in VR with high res headsets like the G2 the GPU will be over stressed in MT, but it's not the case, not even the rendering thread is the bottleneck?! Is this right? with the same settings in 2D I got 120 fps GPU bound in the cockpit VR / Marianas / SP / F/A-18 free flight: how can I say in poline way... well, unusable. Tried to switch off MSAA / forest / scenery details to minimum. Still CPU bound (main thread). It seems that for the current EDGE VR is just became terrible, a stutteting mess. I really hope DLSS and most importantly Vulkan will be close, because right now I'm very dissapointed.

I am one of them. Unfortunately for my system with my settings in VR this MT preview is still far from enough, sorry. No real gains, but some more stutter, missing WOW factor. This is pretty normal right now, this is just the beginning of the road, so definitely no moaning from me. I think the game changer will be the render graph with Vulkan + DLSS. In that way I really hope we'll reach to some heaven when I see outside from the fast moving cockpit and see silky smooth moving trees without hickups. Keep up the good work!

At first I thought reading that 90-120 fps results of yours: "NO WAY". Then I saw your settings. All right, visibility Medium, all shadows and MSAA OFF, water Low. Now I understand, but I'm using G2 not a Quest 2 with much higher resolution (and supposedly much higher picture quality). Interesetingly all my higher settings could mean much higher load on my 3090 but still getting a CPU BOUND state every single time (depending on the map it's main thread or rendering thread). I really really hope that Vulkan will solve this dreading CPU bound because this state is very far from optimal...

I'm really happy to have this update and enjoy the success of others in VR as well, thanks ED! For me personally it's a mixed bag when using the bullt-in instant action SP missions with several airplane types: no breathtaking fps gain (39-40 to 42-43), about 10% maximum, this MT will not give magical 90 fps for VR yet the existing settings has to be adjusted: bye-bye 300% supersampling in OpenXR and using FSR in OpenXR toolkit, because its definitely worse than ST version (even the menu is 75 fps max) more microstutter in MT than in ST, but it's unpredictable; certainly no "buttery smooth" feeling, sorry. There is a good reason, that this is a version got 2.8.3 numbering and not 2.9.0 or even 3.0.0, I know. Eagerly waiting for DLSS and Vulkan for the possible really big boom in VR performance. This version is nice, but this time instead of the WOW feeling I got a meh now. Keep up the good work!

Hurray, it's Friday tomorrow! Another possibility to have some interesting news.

That sounds really good, but these framerates are all for 2D. In VR we're using much higher resolution plus that should be doubled because of the two eyes. The question is that during the rewrite of the new EDGE when MT applied and the graphics and calculations threads are separated are the special techniques already applied with the render graph to avoid the complete re-rendering of the two different images for each eye, or this special VR optimization can only be expected when Vulkan API will be applied somewhere in the future?

All right, then I'll try to be much more on-topic. According to the various leaked information (IronMike's comment on Discord about the allegedly huge gains from the new EDGE and the FPS counter with the text "MT - dx11" on Matt's ARC-210 video on Youtube) some gains can be expected from the hopefully incoming MT version of EDGE but these "leaked"performance data are all over the place. Is there ANY information which can be important to share about the possible performance changes when using VR with high res headsets and strong hardware (CPU / GPU)? I know the mantra that "MT can help mostly in MP and very complex SP missions" but the first newsletter in January which had fine details on these topics in general did not have any preliminary test results. What can someone expect in VR with general SP usage?

Exactly, good point. That's what I said after buying the Apache that no more money from me to ED as long as they do not show some progress. Not even only on the field of development, but the communication of the progress itself. F-15 pre-purchase opened? Not interested, sorry!

After the last delay I'm sure they just returned to their "normal" announcement strategy to say about the patch maximum a day in advance. But it's also pretty much usual to have info only on Wednesday mid day about a patch at the same day (in the evening). So this patch can either be days or weeks away, who knows?

Well, I was wrong. We didn't get ANY news about the forthcoming patch (core systems wise). So sad.

It's Friday, so a newsletter is expected later today. Unfortunately no new info on the patch status tough. I hope at least we'll get some juicy news about the forecoming patch in the newsletter.

This "planned for March" patch became highly anticipated because the highly unusual announcement of it's planned date more than a month early. At that time I already said that the planned 1st March date can surely mean only one thing: no patch in February. I can't remember any occurrence when a patch deployed BEFORE the announced patch date (which is why they usually announce a planned patch date only days before the actual date, not on a distant time scale). This high anticipation raised the expectations and started the speculation that maybe this next patch will include "everything goodies" like MT, DLSS or possibly Vulkan, based on the available infos on the several month long (successful) closed beta testing of MT and the described topics in the first newsletter in January. But in fact ED never confirmed nor denied that ANY of those features could be included in this exact patch. As we don't know how huge this patch will be or how it'll affect core systems we can't speculate on the amount of work and testing it needs to be in the best possible quality before deployment to avoid a bugfest. Personally I always prefer a more thoroughly tested patch than a rushed one. As we don"t know the sheer amount of work and the current progress status we can't imagine that how long the delay could be. And that's my only tiny disappointment in this case: the "we do require a little more time" phrase can be intrepreted as some "more days", "more weeks" on "more months" based on the magnitude of the foreseen amount of work left. Normally I'd bet for "more days" but this time everyhting is unusual. In a theatre when something unforseen happens in the backstage right before starting of the play which can cause more than 1-2 minutes delay someone is sent to the stage to entertain the audience while waiting not just sitting in silence behind the closed curtains... Anyway it's Friday, so I hope at least we'll get a lot of new info in today's newsletter.

Personally I can't because it's already at zero. As there never is any "planned features list" for the next patch announced before the patch itself it's obvious that nothing is preliminary "promised". No one mentioned MT or DLSS for this exact next patch, we don't know ANYTHING about it, so we can't expect more than nothing. It's easy. When the patch will be ready then we can read the patch notes and be happy.

You don't have to guess: they've announced the delay less than a day ago. It's unfortunate, but plans are plans. Now we can only hope that they really just need "a little more time" which is hopefully less than two weeksTM. It would be also nice to have a peek to the planned feature list for this patch at least draft level, but I think it won't happen.

Oh, I forgot that, my bad. I really apologize. Actually I read newsletters occassionally, about 1 per year, that's why I possibly missed that one! That explains everything. I really hope that the hard work of the development team will fruit a stellar new EDGE which we're all waiting for. Thanks for all the newly shared info on core status and feature details!

The ARC-210 radio is a long awaited feature, thanks ED! According to the core: it seems that after some of us mentioned that having some regular report on the progress of core development in the newsletter would be awesome they decided to intentionally not even mention it with a single word, never, ever. Really strange. If this behaviour's purpose to raise the "surprise" factor when the new EDGE will finally arrive then all right, do it. Otherwise it's now becoming really irritating. I never understood why ED's so uncommunicative when the topic is MT, DLSS, Vulkan or other core features which affect all players (VR performance, clouds, dynamic campaign etc.). Anyway, looking for the next wednesday with high hopes.

I'm genuinely amazed, that such a newsletter which have very minimal new information was able to generate more than 150 comments so far. Furthermore all the information included in it was already known, even the starting date of the pre-order period was quasy "known". Not the EA depoyment date, but "only" the pre-order which means several months will pass till the EA launch. So basically people are arguing and flaming about something which is several months away... And not even about any technical details related to the module, but some pricing policy. Unbelievable. But OK, this newsletter was purely dedicated to the F-15, I understand. Now we have 2 days to get the next one, which will hopefully contain ANY information about the core development. You know, 1st March is around the corner...

Pre-order is great and all, but where to rush? Last time with the AH-64 it took 5 months after pre-order to early access launch (ED product), while it was 6 months in case of the F-14 (Heatblur product)...

Nice!