Nvidia’s off-again, on-again virtual GTC 2020 keynote is hitting YouTube this Thursday, and to whip up some excitement, CEO Jensen Huang decided to cook up the ‘world’s largest graphics card’ in his kitchen on an unlisted video on the company’s channel.
While baking graphics cards in the oven is indeed a way to potentially save your dying hardware, Huang’s GPU is an appetizing tease for Nvidia’s next-gen “Ampere” graphics architecture, which is expected to be announced during the event. We expected the reveal to manifest in the form of a massive data center GPU, and Huang’s extra-large serving indeed looks like the “DGX A100.” DGX is the brand for Nvidia’s Tesla-loaded graphics workstations, and Nvidia recently filed a trademark for “DGX A100.”
We’d expect to see Ampere-based consumer GeForce graphics cards later this year, powered by the same underlying graphics technology as the Tesla GPUs inside the DGX A100.
Nvidia hasn’t uttered a peep about its next-gen GPUs beyond this video and a cheeky call to “Get Amped” for the GTC 2020 digital keynote. Ampere’s expected to make the leap to the 7nm transistor process that helped give AMD’s “Navi” Radeon GPUs a huge increase in power efficiency. After losing the 7nm race to AMD, Nvidia must be excited to put its best foot forward before Radeon rolls out more powerful “Big Navi” GPUs later this year.
Both Big Navi and the next-gen PlayStation 5 and Xbox Series X consoles debuting later this year will be built on AMD’s RDNA 2 graphics tech, bringing ray tracing capabilities to the Radeon lineup for the first time. Nvidia (with Microsoft) put ray tracing on the map, so we’d expect the ultra-competitive company to fight tooth and nail with its hardware. Current rumors suggest that Nvidia’s next-gen hardware will include a sizeable boost in real-time ray tracing performance.
Fingers crossed that new-look GeForce GPUs will bolster traditional rendering speeds, too. Despite introducing ray tracing to the masses, the GeForce RTX 20-series found itself rocked by controversy, as most of the consumer graphics cards weren’t any faster than their GTX 10-series predecessors in traditional games (read: 99 percent of games) for the same price. Months later, the RTX “Super” refreshes rectified the situation, and Nvidia’s vastly improved DLSS 2.0 is now firing on all cylinders, too.
Hopefully, Nvidia learned its lesson after RTX 20-series got off to a slower-than-expected sales start. Even in data center form, we can’t wait to see what’s next on Thursday.