I wonder how we might be able to run a Pimax 12K. What hardware may be required.
Let us take HLAlyx, fitted resolution for announced Pimax 12K.
To have sufficient frame rates and low or no framedrops, what hardware would be required?
I mean, without upscaling, which seems to be a trick for me, to avoid mentioning what really is required.
We only can guess, so I will do:
We need something like 2x NVIDIA RTX A5000 or successor, coupled via NVlink. This may grant enough compute power AND enough graphics memory. I do not see an equivalent besides that.
I hope, Pimax considers these facts to enable real 12K aka 6K stereoscopic vision.
Am I wrong?
1 Like
DFR actually makes the 12k, on average, more performant than the 8k X is. They said something like gains of 37% on average across 10+ games using 3080s to test supposedly. Youâll probably need less powerful hardware than you think for games that support variable rate shading, which is a decent amount, if that holds true.
You mean a solution for kind of a LAN Party where you may use four Headsets in parallel, I guess, or kind of an Arcade Version? 
Ah, I got it, it is an Internet Coffee solution, isn´t it? 
Impressive, but isn´t it a little bit overpowered for home use? I may need extra power supply installation to drive it. 
I believe Dynamic Foveated Rendering (DFR) is kind of a Remedy to reduce compute power requirement, if well implemented. In combination with upscaling it may help a lot with existing consumer hardware like RTX 3080/90.
On the other hand, still it is kind of a trick, which can be recognized, therefore immersion lowers.
From a bionic aspect, I hope we will see true 90°-100° stereoscopic overlap, for a better immersion.
Of course it is all about the UVoV (Useful field of view), not just the technical FoV, e.g. measured from âedge to edgeâ, but true horizontal like our bionic sense ability about 220°
And of course, we would like to see some 90Hz frame rate or even better.
Knowing that, I still fear a Nvidia Geforce RTX 3090 may not be enough compute powerâŚ
DFR works on the same principle our actual eyes do in real life. When properly implemented, you wonât really notice the difference. Regardless of anything being a âtrickâ or not, if you donât notice the difference, then does it really matter?
You are right, it does not matter if DFR is properly implemented.
Maybe we will see certain iterations of the software, until we may agree it is at least acceptable, or hopefully we will see a proper successful development.
I am curious if Apple will catch up or even outpace Pimax. Or Facebook (âmetaâ) will make it. They always invest heavily to kick out any competitor. I believe we need Pimax as a technology leader.
So, we stay tuned.
Just stumbled over some news:
A report about PCIE 6 refers to PCIE 5 to be replaced in 2023 (original source PCI SIG).
Together with the news about the delay of the GeForce RTX 3090 Ti this year (2022) (original source: see TweakTown article), which will base on PCIE 5, we may see a possible realistic consumer setup end of 2023, which may be able to run the PIMAX 12KX with a certain decent resolution and frame rates. Still: it would require at least a new mainboard (PCIE 5) and an outstanding expensive consumer graphic card, because of being the new NVIDIA flagship (nothing available or announced from AMD as far as I know).
It would require a new AlderLake platform (Intel) or expected Zen 4 (AMD).
As we may see further delays on all sides, a realistic launch may be end of 2023 (maybe paper lounch) and availability of required components may be 2023.
Then start counting the bucks (Dollars/EurosâŚ) you need to invest
just my thoughts.
Due to the (partly aspherical) lenses and the higher ppd /native resolution, it will look better than many other HMDs even without supersampling. So even with the settings we are running on the 8kx now.
Nobody forces us to use the possible maximum.
I think you are severely overestimating what will be needed to make âgoodâ use of the 12k. Factoring in DFR itâll already be perfectly possibly to make âgoodâ usage of it with high end current hardware. Maybe not max out every in game setting, but run at native res/90hz and fairly good settings. At worst maybe lose a little FoV. Things like the eye tracking enable foveated transport which also reduces the bandwidth needed to run the headset.
12k will need a 5090 card minimum to run something like msfs 2020 at full res full fov and high refresh rate.
I am not really fully convinced if DFR is the remedy here. First, I would like to see that Pimax can handle eye tracking. Then⌠I would like to see the results of DFR implementing first. May be ingenious on paper but maybe then not acceptable in reality.
For me, I do not want to lose any FoV. To buy Pimax because of te outstanding FOV, and then a loss of VoV because of software inconveniences or insufficient hardware? No way. I could directly switch to StarVR or Varjo or other headsets available that time with âlow FOVâ.
So, I want to see ⌠Anyway we have to wait for another year at least, I estimate.
I see this as Pimax 8KX Plus or Pro. Why? Because nobody is going to use the 200 degree vertical FOV anywayâŚfor example, in the 8KX I never used 170 degree due to the hardware limitation and I have RTX 3090. Unless they have DFR working for the game (and some games might not even support it), you will have to get 4090 to run this thing. In the end, the only good thing about this device isnât because of its high resolution (the 8KX is good enough imo), itâs because of the auto IPD adjustmentâŚwhich they screwed it up mechanically with the 8KX. I see them as trying to remedy the situation by giving us the discount toward 12K which seems fair. Other features like wireless, ect., I do not have high interest since I use it mainly for sim racing.
ah⌠seems to be different for me. I like Aircar, where the FoV of 8KX works pretty well (with an RTX 3090).
both horizontal and vertical FoV count, if you are spiraling through the air. Just try it, if you have no issues with motion sickness, like me. It´s just another user scenario, as worth as yours. As well as many others, me included would like to go literally wireless.
Have a look at my inital post, I fear, a single card (RTX 4090) still may not be sufficient to run it properly with these adverdised / announced features. I still believe âWe need something like 2x NVIDIA RTX A5000 or successor, coupled via NVlink.â
we could also suppose that PIMAX is announcing an incredible headset far ahead form others just to get attention of Apple and get taken over (sold) by a big company (Apple or Google or Microsoft).
PIMAX is still small enough but a serious contender in VR headset to be credible while still affordable.
âŚthat PIMAX is announcing an incredible headset far ahead form othersâŚ
You mean like Meta did these days 
20° FoV, even with retina resolution, I am not convinced.
I believe in Pimax as a proud company being able to evolve, grow, prosper and make big business without being taken over by GAFA+M.
PIMAX12K seams reachable, reasonable for now. Hope they can deliver in time.
If not, there is Apple with its own M1 M2+ultra whatever silicon, which really could be a game changer at least for mobile/wireless applications and AR integration.
There is room for improvement for a better VR/AR cinema, which is the consumer version, where VR/AR meeting and augmented services is the business version.
With the latest rumors about Nvidias 4090 where they seem to find performance doubled compared to 3090, now there is a fair chance to run a Pimax 12KX, maybe with above mentioned limitations.