The new golden age for chip mobile? (SoC)

John117-

n00b
Joined
Jun 5, 2021
Messages
22
So the smartphones were the best place for developing better the chip mobile. Thanks they, we had a increasing of performances for the mobile computing. After around 15 years of the first release smartphone, we have SoC like A14 that it has 6 core and 6GB of RAM, for example.

Meanwhile, the state of art for VR is XR2 that it has 8 cores, it can manage 7 cameras.

In my opinion the VR/AR mobile computing are arriving in the best moment for push the development of chip mobile. The smartphones aren't anymore the best platform for pushing the performance of SoC. I'm not belittling these products, but after 14 years of continued increase performance, it's arrived the time of change platform. In other topics I expressed my doubts about abbandoned of PCVR for developing before on the standalone segment. I have still these doubts but it's not this topic for speak it.

So, the standalone segment will be the landmark for engineers for developing the performance of SoC. I explain my opinion: today we have the right power into the smartphones.Anyone can buy a good smartphone for accesible price. We're facing a technological stalemate about smartphone, now the focus is the camera in many cases. The engineers doesn't have motivation to push performance on smartphones, the VR/AR mobile can push the engineers for developing the performances of the SoCs.

We need of more power for have better experience in standalone segment. Imagine a hmd 4K for eye, Eye Tracking, 120 Hz. There's need of more power, obviously for manage it.

Another point very important in my mind is heat of the SoC and the probability of Thermal throttling. Beyond of the manage of battery.
The manage of heat is very important for the hmd standalone. You can imagine the problems without a good system of cooling. More power without problems of thermal throttling is the main goal, in my opinion obviously. The development of better batteries for mobile devices is another topic, that maybe I will discuss. We needs of better batteries for durability and SoC with amazing efficient.

Let me know your opinion, thank you :)
 
Interesting, I haven't made that association in my mind previously.
To be honest, I kind of ignored the VR technology simply to it not being 'my thing'.

And I was convinced that those are two separate groups of devices, with the mobile ones also handling IO, power management, audio...
In my mind the VR hardware was more akin to a GPU. Am I wrong?
 
Standalone VR is nowhere near being in a position to drive silicon design, just on a pure numbers level. Like, total VR headset sales in the past year -- standalone & tethered combined -- only amount to 6.1 million units (~5 million of that being the Quest 2). For perspective, Apple sells that many iPhones in two bad weeks (or one good week). No one's designing bleeding-edge silicon for consumer electronics with sales that low -- sales which would tank hard the instant hardware costs ticked above $500.

And no, the XR2 isn't evidence to the contrary. It's just a Snapdragon 865 with a camera tracking block duct taped to the side. That's as close to no-effort as Qualcomm could get without just deciding not to bother.

Maybe, if the VR ecosystem digs itself out of its current malaise (PSVR2, save us!), and Facebook manages to double uptake of their standalone headsets (while continuing to subsidize the everliving hell out of them), then they might be in a position to dictate upcoming silicon.


Interesting, I haven't made that association in my mind previously.
To be honest, I kind of ignored the VR technology simply to it not being 'my thing'.

And I was convinced that those are two separate groups of devices, with the mobile ones also handling IO, power management, audio...
In my mind the VR hardware was more akin to a GPU. Am I wrong?
Standalone headsets are basically the guts of a phone grafted to VR optics. Top of the line is on the verge of Xbone/PS4 performance, which is just enough for tolerably convincing VR performance in the hands of decent developers.
 
Last edited:
Standalone VR is nowhere near being in a position to drive silicon design, just on a pure numbers level. Like, total VR headset sales in the past year -- standalone & tethered combined -- only amount to 6.1 million units (~5 million of that being the Quest 2). For perspective, Apple sells that many iPhones in two bad weeks (or one good week). No one's designing bleeding-edge silicon for consumer electronics with sales that low -- sales which would tank hard the instant hardware costs ticked above $500.

And no, the XR2 isn't evidence to the contrary. It's just a Snapdragon 865 with a camera tracking block duct taped to the side. That's as close to no-effort as Qualcomm could get without just deciding not to bother.

Maybe, if the VR ecosystem digs itself out of its current malaise (PSVR2, save us!), and Facebook manages to double uptake of their standalone headsets (while continuing to subsidize the everliving hell out of them), then they might be in a position to dictate upcoming silicon.



Standalone headsets are basically the guts of a phone grafted to VR optics. Top of the line is on the verge of Xbone/PS4 performance, which is just enough for tolerably convincing VR performance in the hands of decent developers.
Your opinion is interesting, but I'm not agree. I'm so sorry.

Yes, XR2 is almost a 865. But It's the state of art for mobile VR, seriously. Do you remember GearVR or Oculus Go?

What's the problem with Smartphones? In my opinion the sales are very important, obviously, but we're at the end of the power race about the smartphones. There isn't more needs of more power on smartphone. Obviously It's not possible that Qualcomm leave snapdragon series for smartphone for developing only VR SoC.
My point is another: Samsung could make a customize a SoC Exynos for VR standalone, Facebook in the next years could make customize SoC for Quest Pro and another Quest, Apple with his HMD could make a SoC personalized or Valve. In conclusion, the SoCs are very powerless compared at GPU/CPU, obviously. There's the problem of management heat, one of the point more important for HMD Standalone. We will see new advance in this aspect thanks to VR Standalone.
 
Your opinion is interesting, but I'm not agree. I'm so sorry.

Yes, XR2 is almost a 865. But It's the state of art for mobile VR, seriously. Do you remember GearVR or Oculus Go?

What's the problem with Smartphones? In my opinion the sales are very important, obviously, but we're at the end of the power race about the smartphones. There isn't more needs of more power on smartphone. Obviously It's not possible that Qualcomm leave snapdragon series for smartphone for developing only VR SoC.
My point is another: Samsung could make a customize a SoC Exynos for VR standalone, Facebook in the next years could make customize SoC for Quest Pro and another Quest, Apple with his HMD could make a SoC personalized or Valve. In conclusion, the SoCs are very powerless compared at GPU/CPU, obviously. There's the problem of management heat, one of the point more important for HMD Standalone. We will see new advance in this aspect thanks to VR Standalone.

"State of the art for mobile VR" doesn't mean anything. It's literally nothing special next to flagship phone silicon -- because it is phone silicon. It was immediately surpassed in performance by the Snapdragon 888. And neither holds a candle next to Apple's M1.

And yeah, I remember GearVR and Oculus Go (and Google Daydream). They were 3DoF toys -- underpowered, overpriced, moderately nauseating, and entirely awful. They were vaguely interesting for a weekend, after which they collected dust in a box.

I have doubts that the Samsung/AMD deal will result in competitive hardware any time soon. Key, for me, is that AMD has (to my knowledge) no experience in this ultra-low-power space. My gut says we'll see a number of generations following the Tegra product timeline -- moderately performant silicon, but with power efficiency uninspiring enough to render them largely undesirable. Which pretty well describes Samsung's ill-fated experiment in custom CPU cores, before they went back to slinging ARM IP blocks.

Meanwhile, Facebook's dabblings in ARM SoCs are limited to datacenters. And if Apple ever dropped a standalone headset, its performance would lie in the shadow of whatever Macbook its SoC was derived from.

In conclusion, your conclusion is baseless. You can't claim that mobile silicon is underpowered when said silicon is absolutely dominating on performance per watt, and Apple is actually using said mobile silicon as the basis for a wildly-successful revamp of their entire computing line.

...And that last point indicates where you should actually be looking for the bleeding-edge advancements in low-power SoCs -- ARM-based computing. Sadly, we'll never get Apple silicon outside of Apple products, but if Microsoft can avoid screwing up their next swing at Windows on ARM, we'll start seeing demand for highly-performant SoCs in volume, which will greatly benefit future standalone VR headsets.
 
"State of the art for mobile VR" doesn't mean anything. It's literally nothing special next to flagship phone silicon -- because it is phone silicon. It was immediately surpassed in performance by the Snapdragon 888. And neither holds a candle next to Apple's M1.

And yeah, I remember GearVR and Oculus Go (and Google Daydream). They were 3DoF toys -- underpowered, overpriced, moderately nauseating, and entirely awful. They were vaguely interesting for a weekend, after which they collected dust in a box.

I have doubts that the Samsung/AMD deal will result in competitive hardware any time soon. Key, for me, is that AMD has (to my knowledge) no experience in this ultra-low-power space. My gut says we'll see a number of generations following the Tegra product timeline -- moderately performant silicon, but with power efficiency uninspiring enough to render them largely undesirable. Which pretty well describes Samsung's ill-fated experiment in custom CPU cores, before they went back to slinging ARM IP blocks.
Samsung/AMD deal IMHO has to be based on something. Two tech giants in these times and you say "meh"? Bro... come on.
Something is up and we are fusing micro cores with high performance GPU tech. How this will play out? Who knows ya?
But it aint the BS you are spouting for sure!
Personally my future vision is a "Steam-deck" like device working some magic to a high powered GPU.
 
Samsung/AMD deal IMHO has to be based on something. Two tech giants in these times and you say "meh"? Bro... come on.
Something is up and we are fusing micro cores with high performance GPU tech. How this will play out? Who knows ya?
But it aint the BS you are spouting for sure!
Personally my future vision is a "Steam-deck" like device working some magic to a high powered GPU.

They joined together to produce the Exynos 2200 chip for the new high end Samsung phone. But, I believe the majority of their high end phones will use Qualcomm hardware.
 
They joined together to produce the Exynos 2200 chip for the new high end Samsung phone. But, I believe the majority of their high end phones will use Qualcomm hardware.
Yes they aren't going all in atm, but I feel it would be a wasted endeavor if they didn't have much bigger plans going forward.
 
Yes, XR2 is almost a 865. But It's the state of art for mobile VR, seriously. Do you remember GearVR or Oculus Go?
Not sure the relevance according to the op subject, if something was developed because of smarthphone and happen to be used by a vr system, it is still an clear example of the : smartphones were the best place for developing better the chip mobile

VR do seem a bit too niche for that, one giant SOC development was the recent Apple PC line for a very recent example and obviously phone still going on (for some reason not sure what will be done with stronger phone), but that, regular next generation of consoles and intelligent cars are the more obvious suspects for what will be done, with VR using them instead of pushing them being what is more likely to happen.
 
I think that we're talking about of different vision for the future of SoCs. The smartphone are the main market because they're more popular and economic. The market of smartphone is very big, so it's normal that the architect of chip, design first for they. My vision is for the future. In this moment there's a dead point for the power of smartphone. They doesn't needs of more power, not like headset VR/AR. I say that thanks to smartphones we have standalone headset today. But for the future, companies like Facebook/Apple and other for example Microsoft (if they enter in this market), Samsung or HTC (but htc eeds of money for custom chip) could make a custom chip for VR/AR headset. My idea is for the future of design about SoCs is that VR/AR could push the actual limit.
 
I can see Apple designing their VR/MR device/headset to be powered by an iPhone and therefore a mobile SOC with VR needs. The biggest issue to custom silicon design is volume. Oculus has some decent volume but they would need Qualcomm or ARM to make a special arrangement for VR use. Microsoft could do full design but they seem to have cooled on MR. It would be neat so see a new balls to the wall VR SOC but not sure who would do it - maybe Nvidia if they buy ARM.
 
Back
Top