But one paper suggests that we’d be better off betting on advances in local computing, rather than having to use remote computing over unreliable networks see Section 2 . Cloud-based rendering and video streaming is a compelling idea, but it also greatly increases the amount of low-latency data that needs to be delivered. As mentioned earlier, gaming content aims for at least 60 frames per second more than twice the video standard, and even 90-120 frames per second, ideally with 2K to 4K clarity. Delivering video streams reliably to everyone who wants to participate in the Metaverse while keeping latency low is really hard. down; we don’t work to get the electricity we need every day, nor do we get it as quickly as we need it.
Even at ultra-low latency, streaming vs. processing locally AR data uganda mobile database doesn’t make a lot of sense given the speed at which input data is received i.e., the speed of light, at a distance of just a few feet. Given the intensive computational requirements of AR, our personalmobile devices are likely able to do a “good enough” job for most real-time renderings.
So far, remote computing hasn’t proven to be more efficient for rendering, either. That’s because cloud-based GPUs don’t produce general-purpose rendering “power.” Instead, they lock down compute instances: a single GPU, remote or local, supports rendering for only a single user. No one has figured out how to efficiently, economically, and at modern expectations for resolution and frame rates distribute its rendering power across multiple users, just as power plans distribute electricity across multiple homes.