Wednesday March 20, 2019

Nvidia Skips Ampere at GTC 2019

Several news outlets seems to think Nvidia's GTC presentation was relatively longwinded and unexciting this year. The three-hour keynote reportedly featured some software announcements and a low power Jetson board, among other things, but didn't feature the 7nm Ampere GPUs many were expecting. EE Times says that the "unspoken message" at the presentation was that "Nvidia doesn't need to pre-announce a new and faster chip because it owns that software stack and channel today," and the emphasis on CUDA seemed to really drive that point home.

However, in one of the more exciting parts of the presentation, Nvidia did highlight the Q2VKPT project we covered earlier this year. Nvidia's CEO seemed quite excited about the introduction of raytracing to Quake II, and they showed off some of the project's gameplay, which you can see below:

Presaging that future, Nvidia's chief scientist, Bill Dally, told reporters about a research project in optical chip-to-chip links. It targets throughput in terabits/second while drawing 2 picojoules/bit/s. In an initial implementation, 32 wavelengths will run at 12.5 Gbits/s each, with a move to 64 wavelengths doubling bandwidth in a follow-up generation. Dally predicted that copper links will start run out of gas as data rates approach 100 Gbits/s, already on many roadmaps for network switches. Progress in more power-efficient laser sources and ring resonators will enable the long-predicted shift, he said. If the future evolves as he believes, bleeding-edge GPUs may continue to skip an appearance at some Nvidia events. Attendees will have to hope that as the interconnects speed up, the keynotes don't get even longer.

Discussion