Intel issued its first statement on the Metaverse on Tuesday. This is the first public recognition of this sometimes vague future of computers, promising an always-connected virtual world that coexists with the physical world. However, while chipmakers are optimistic about the possibilities of the Metaverse in abstract terms, Intel raises a central issue in realizing all the Metaverse ambitions.
"Metaverse has the potential to become the next major computing platform after the World Wide Web and Mobile," begins an editorial by Raja Koduri, Senior Vice President and Head of Intel's Accelerated Computing Systems and Graphics Group. But Koduri quickly pours cold water into the idea that the Metaverse is right there. "Our computer, storage, and network infrastructure isn't enough today to make this vision possible," he writes. The bottom line is that Oguri doesn't even think we're nearby. He states that it will need to improve performance by a factor of 1,000 over current collective computing power.
Much of the Metaverse hype is based on what happens there, such as virtual reality conferences, digital concerts, and of course blockchain and NFT-based integration. The future of virtual and augmented reality headsets, whether Meta (formerly Oculus) Quest products or Apple's long-known headsets, is also exciting.
But the actual
But the actual components of the Metaverse are software and virtual space (of course, the struggle of its own because today's digital world is so self-contained) and people wear to get there. Not even headsets or gadgets. .. Metaverse positions the future of technology as computers and servers running vast shared virtual worlds. And that's where Intel does its biggest reality check. Computers today aren't powerful enough to make those dreams come true. They aren't even around.
Today's computers don't have enough power in the Metaverse On the one hand, the statements here are almost ridiculously obvious. Meta's flagship VR room, Human Worlds, has up to 20 participants in a room for a simple Roblox-style animated world. VR's state-of-the-art technology costs thousands of dollars for PC gaming hardware, but it has many drawbacks (for example, even the largest traditional video games that don't meet the additional requirements of VR, such as Fortnite and Battlefield 2042. , Can only be dealt with) Up to 100-128 players can participate at the same time.
As Koduri points out in an editorial, today's technology can't even put two people in a very detailed virtual environment. "Think of what it takes to put two people in a social environment in a fully virtualized environment. Attractive and detailed avatars of realistic clothes, hair and skin tones-all rendered in real time and sensors. Data, real 3D objects, gesture-based, audio, etc. Ultra-high bandwidth and ultra-low latency data transmission. And persistent environment that can contain both real and simulated elements. Model. Requires far more computing infrastructure.
Of course, Intel
Of course, Intel is also legitimately interested in saying that it needs more and better computers and servers. By the way, Intel manufactures CPUs (and soon GPUs) for consumer devices and data centers, and Metaverse (the hottest buzzword technology of the future) literally increases computing power by a factor of 1,000. When you need it, it's only good for your business. It's no coincidence that Intel explicitly mentions both client computing and cloud processors and graphics products in the Metaverse briefing.
The problem, though, is that even Intel doesn't believe that hardware alone will multiply it by 1,000. As Koduri explained in an interview with Quartz, "I believe that the standard curve of Moore's Law will only grow about eight or ten times over the next five years." (Moore's Law generally defines doubling computing power every two years, which is consistent with Koduri's 8-10x growth forecast.)
Instead, Koduri
Instead, Koduri uses an algorithm. We are optimistic that the gap will be closed by improving the software. Intel is already using it for deep linking technology and XeSS supersampling, which will be introduced on ArcGPU early next year, such as neural networks using machine learning and AI-enhanced computing technology. However, this is a big challenge. In addition to the growth provided by existing hardware roadmaps, Intel relies on algorithms or AI to increase computing power hundreds of times (or more).
Koduri
Koduri said in the same Quartz interview that more than just software and algorithm improvements are needed to fill the ambitious five-year gap he set. They are important to mitigate the increased energy consumption that results from trying to cause brute force problems. This is what he compared with the current problems of cryptocurrency mining today.
Hardware improvements are not enough It's easy to wave and say that the software fills the gap left by the hardware (especially for companies like Intel, which mainly manufactures hardware). Many big tech companies have embraced the idea that AI and machine learning solve computing problems, from improving smartphone cameras to delivering high-end gaming graphics, and you'll want to think it's possible. .. But it still seems like a big job that depends 100 times more on tomorrow's computing. This only predicts a 10x jump with just a hardware improvement.
But the fact that Intel is now pondering all of this and identifying the problem is a reassuring sign. Riding hype is easy and you can come up with great ideas for selling NFTs that follow you from place to place in a variety of games and virtual environments. Strengthening the server infrastructure and working to reduce latency is less attractive. But, as Intel's presentation shows, there's more basic work to be done in the coming years to pave the way if the Metaverse hits science fiction ambitions.
