Is VR dead?

[RCHeliguy]

If that is correct (guess it is) then it means a lot more than just a number.
Mainly because smaller nodes/finer resolution means (generally) better/lower power use.

We can hope! If this were directly linear, this would mean 20% less power, but I'm sure it's more complex than that. Typically smaller also means faster and the Nvidia card clock speeds are still well below the 5GHz barrier that Intel has been stuck at for a while.

 

[BrunoBæ]

Yeah maybe I should have explained myself better.
Because until now then using finer/lower node size have almost allways been combined with higher speed (Ghz) so the end result have been the same or often higher power demands.

 

[RCHeliguy]

Recent benchmarks are suggesting that the 40 series will have 2.5 X the raw compute power of the 30 series and that the 4090 will have 100TFLOPS of 32bit floating point computation.

In addition to using 4nm process manufacturing, they may also start using hopper technology in the 40 series. Looks like it should be something very special.

However other rumors have suggested that the Index 2 won't be released until 2023

 

[BrunoBæ]

I just found this description on nVidias site.
And I have to admit that if I had found such vague uberhyped text on as example Pimax site I would have laughed and ruled this out as the mandatory untrustworthy Pimax bull.
But I have also to admit that eventhough this still is pretty vague then maybe nVidia has actually something up in the sleve.
Hehe eventhough I still dont believe the 4000 series will be more than max 30% faster than the 3000 series.
Hopefully Im wrong

"The NVIDIA Grace Hopper Superchip leverages the flexibility of the Arm architecture to create a CPU and server architecture designed from the ground up for accelerated computing. H100 is paired to the NVIDIA Grace CPU with the ultra-fast NVIDIA chip-to-chip interconnect, delivering 900 GB/s of total bandwidth, 7x faster than PCIe Gen5. This innovative design delivers up to 30x higher aggregate bandwidth compared to today’s fastest servers and up to 10x higher performance for applications using terabytes of data."

 

[RCHeliguy]

Hehe eventhough I still dont believe the 4000 series will be more than max 30% faster than the 3000 series.
Hopefully Im wrong

Given they have dropped from 8nm to 4nm and dramatically increased the power requirements, I don't think something small like a 30% increase is remotely possible.

I think the doubling of performance is likely very close to the truth. Now I just need a headset to drive with that.

 

[RedBull48]

Given they have dropped from 8nm to 4nm and dramatically increased the power requirements, I don't think something small like a 30% increase is remotely possible.

I think the doubling of performance is likely very close to the truth. Now I just need a headset to drive with that.

Are you currently using the valve index? Are you considering anything like the Varjo Aero?

 

[RCHeliguy]

Are you currently using the valve index? Are you considering anything like the Varjo Aero?

I'm using the Index with 2080Ti for a while now and I'm going to wait for the next Valve headset at this point. It just may be a long wait

 

[RedBull48]

Yea, I am really looking forward to the index 2 as well. But I am also still waiting for my steam deck that I reserved almost a year ago . I am thinking of getting a reverb G2 to upgrade my rift s in the immediate future but I am on the fence.

 

[wajdi]

Yea, I am really looking forward to the index 2 as well. But I am also still waiting for my steam deck that I reserved almost a year ago . I am thinking of getting a reverb G2 to upgrade my rift s in the immediate future but I am on the fence.

Get the G2 Reverb . I had the oculus S them upgraded to the G2 and it is as a big step in terms of clarity and image quality . Get the FOV mod gasket and it’s even better. I wait for the Index 2 to but I suspect we won’t see it untile the end of 2023.

 

[Deleted member 197115]

Remember 3080 announcement?

 

[Bjarne Hansen]

You can pretty much say that about every Nvidia GPU generational update.
But hey it works for some!
Here there are al least one that believes it

If I recall right only the Gtx 7xx to Gtx 9xx upgrade was a big performance jump where the 970 was better than the 780ti and only when OC which they did well.

 

[BrightCandle]

Its the power consumption rumours that really bother me, I remember the 580! Hot, loud and expensive that is what you get with high power consumption, ignoring the fact that it also costs substantially more to even run the card too. There is simply no way around the physics of cooling and noise is something I personally hate.

 

[blekenbleu]

I had missed that some premium smartphones already have LIDAR
LIDAR scanning applications for iOS and Android

 

[wajdi]

 

[RCHeliguy]

 

[RCHeliguy]

If 40 series were released this July, the new power supplies will need to be released.

 

[RCHeliguy]

So 4090 is now only 450W instead of 600W ?
And 4090 = 2 x 3090's for performance.

But this time they are backing that up. This isn't in terms of some new way to measuring things. They've nearly doubled the number of general purpose cores which is a real way to double the performance.

I'm guessing that 450W instead of 600W is because of their dropping to 4nm process.

 

[wajdi]

 

[blekenbleu]

I'm guessing that 450W instead of 600W is because of their dropping to 4nm process.

At least when I was involved years ago, +/-20% power estimates were typical at tape-out,
because voltage would be tweaked +/-10% to get yield on a new geometry generation.
Engineering changes between first test wafer and production improve yields
and reduce power. Estimates for first designs at these days' smaller geometries
may want even more conservative estimates.

 

[BrunoBæ]

If these values also holds on other games then the 4000 sewries IS about doubling the performance.
Just waiting to hear about prices, power demands and supply.
Maybe my good old thrusty 1060 will have to retire - some time in the future

CatsAreTheWorstDogs: Ultra Nightmare Quality