About 60-70% performance improved on RTX2080Ti


#84

Yes, exactly, that’s the same thing

The feature was well described to the testers in the closed forum. I think all testers participated quite a bit in that discussion.


#85

Ouch… Do you mean “…mostly kids who need simple explanations.”?

EDIT: Bitching? Excuse me? You obviously didn’t read my first post which was complimenting Pimax for what I thought was a new improvement.


#86

And how many of them provided performance comparison and explained it to us? None. This is what the topic about, to explain it.


#87

Well they were under NDA back then. Yes, this topic is to explain it to the public, but it seems a bit dodgy the way it’s presented here, like they’ve made a recent breakthrough that’s now giving 60-70% performance improvement. That’s not true. If they just wanted to explain the options in pitools, then they should have also explained the other options and probably named it something like ‘pitool explained’ or something.

Anyway, no biggie, but it’s good that it’s cleared up now I think.


#88

Only Pimax 8K or 5K+ too? :cold_sweat:


#89

Sad to report that X-plane 11 and Aerofly FS 2 have to use parallel projection mode.


#90

[quote=“risa2000, post:81, topic:9072”]
LoneTech:
“The non-coplanar mode can be accelerated with simultaneous multiview rendering on the 20 series. The 10 series was advertised with the exact same feature, but it was crippled to views that were not only coplanar but only offset along the X axis.”

Could you share any reference to those statements? Until now I thought that the only difference between GTX and RTX in multi-view rendering was the number of views it could do simultaneously, the former only two, the latter four, but no other restrictions. If, as you say, GTX support is limited only to co-planar projection planes then it could significantly sway my decision about which card to get.[/quote]
He is correct.

Multi-View Rendering (MVR)

MVR powerfully extends Pascal’s Single Pass Stereo (SPS). While SPS allowed rendering of two views that were common except for an X offset, MVR allows rendering of multiple views in a single pass even if the views are based on totally different origin positions or view directions. Access is via a simple programming model in which the compiler automatically factors out view independent code, while identifying view-dependent attributes for optimal execution.


#91

For what reason Does pimax have to lie ? Last I checked they’re not in the graphics card selling business .


#92

But good for us that chose 8k yay!


#93

The gain of 30% (10-30%) in the 2080 ti is the same as in any other monitor game or VR headset, so Pimax does not use too much of those cards yet.

This “Compatible with parallel projections” check box has been known to testers for quite some time. It would have to be activated for games in which the image appears duplicated in both eyes, which happens in many of the games.

The Chek box “Hidden Area Mask” consumes quite a lot and only serves to fix artifacts that appear in a handful of games, I think that the artifacts were due to that some games, I suppose that to increase the performance, did not draw everything that was seen with the FOV of the Pimax, remaining black zones or with residual images.

The testers already recommended to Pimax that they would have to detect when each of the options were necessary and activate them automatically without the user having to do anything, or at least save the options for each game.


#94

Several additional improved solutions are to be done.

  1. we just found some extra rendered vertical FOV which exceed the real vertical FOV. that waste about ~10% of current total rendering pixels. we will try to reduce the rendered vertical FOV but keep some margin and not impact the real visual FOV. Hope this helps us improve the whole performance 4%~6%.

  2. Nivida keeps improving their driver for RTX2080/RTX2080Ti. This might help a bit. 3%-6%?

  3. We’re developing ASW. ASW will help a lot for the real VR experience while the FPS drop off.

  4. 72Hz or 65Hz optional mode will do helps, might another ~8-10%?

Planned to be supported by Pimax exclusive SDK in the future.
fix foveated rendering and other new techs from VR works or LiquidVR.
60Hz*2 for BrainWrap in the future, we’re trying.

We believe we put more effort into the software fine-tuning and optimization, more chance of performance improvement we can achieve. Since Pimax 8K/5K+ and RTX2080/2080Ti are newly born, they definitely have some margin to be optimized.​


Wouldn’t fixed Foveated Rendering help a lot?
#95

@PimaxVR could you please clarify if this (number 1) improvement applies to any graphics card, or only to RTX 2080 (Ti)?

I understand that point 2)

just refers to the overall performance increase between GTX 1080 Ti and RTX 2080 Ti, but the applicability of point 1) is not clear from your post.


#96

A criticism of Pimax and the 8K especially is that there aren’t GPUs powerful enough to run them yet. It has been speculated that even the 2080Ti increase would be 20-30% and even that’s not enough. So suggesting that the 2080Ti delivers a whopping 70% boost over 1080Ti suddenly quashes this argument and makes the Pimax (especially 8K) a viable option. So they may not be selling 2080Tis but they are very dependent on it being a success in order to shift their own units.

Now I don’t necessarily agree with any of the negativity in the first place but their ambiguous claims definitely suit them.


#97

And then telling the three testers to confirm what they found ?Seems like a foolish Bluff let’s see

I just have to have the best maybe I’m trying to convince my self to pickup the 2080Ti that I pre-ordered , it’s in store tomorrow Decisions decisions .


#98

We have achieved ~35% performance improvement by implementing a new rendering algorithm, but it’s not backward compatible. it requires Unreal 4.18 or above version, and Oculus SDK 1.17 or above. But almost every Unity games are OK.
@PimaxVR could you please clarify if this (number 1) improvement applies to any graphics card, or only to RTX 2080 (Ti)?

The new rendering algorithm gets improvement for all video cards. but might gain more for the new card, but less for the older card. (only roughly estimated just for reference, not the precise statement or even a wrong estimation due to lack of enough test). 1080ti seems close to 30% 2080ti seems 30~35%. but differ from different games behavior.


#99

Excellent. thank you.
Do you happen to know how many HMD`s will be in the next batch ?


#100

So the new minimum graphics card is a nVidia RTX 2080 Ti for both 8K and 5K+?
I think that Pimax would have to work much more on the performance in the 1080 and 1080 Ti cards instead of expecting that the lack of performance has to be paid by the user of the pimax, buying the most over priced cards of nVidia, and changing much of their PC too so that the processor and memory do not make bottlenecks.

As for the software changes of nVidia and AMD, they can take many months to be useful and even some of the technologies can end up being abandoned or be useless for Pimax.


#101

Hello Pimax,

Please do not say that the engine for IL-2 Sturmovik: Battle of Stalingrad is powered by Unreal or Unity. It is NOT in any way powered by Unreal or Unity. Our Digital Warfare Engine is a custom graphics and physics engine built from scratch and improved over many years now. We are not an off-the-shelf engine like Unreal or Unity and therefore any such changes to how VR is rendered is potentially a lot of difficult work. Even so, our current VR implementation is arguably one of the best out there when you consider how large our environment is and certainly the best when it comes to combat flight-sims. We will support VR technology whenever feasible for our small team.

Regards,

Jason Williams
Executive Producer
IL-2 Sturmovik: Great Battle Series


#102

If the rest of the testers take a look a their spreadsheet of games, it sais that is
Digital Warfare Engine original from 777 Studios.

I was looking for the engines of each game, to see if the problems that appeared could be due to any of them (not this case).

Where did you see it said to be Unreal or Unity? Ah ok, the first entry.


#103

@PimaxVR Thanks for the update, Also, can you confirm the color and black level adjustments for the Pimax 5K+ are being made, so that the 5k+ has the same color and black levels than the 8K, as mentioned in this post ?

Accordingly, we tuned a special gamma setting for Pimax 8K under 1.75 rendering quality value, it can enhance the sharpness but drop off some brightness.

About the ‘special gamma setting’ for the 8K, can it be adjusted manually by the users in Pitool ? what if some users don’t want to trade sharpness for brightness in certain cases ?

And are you aware of the distortions on the 8K when turning very fast or making rapid motions with the headset and do you think you can you do something about it ? (MRTV did mention this several times in his videos)