Become a fan of Slashdot on Facebook


Forgot your password?

Submission + - AMD Next Gen Navi GPU Architecture To Employ AI-Dedicated Circuitry ( 1

MojoKid writes: Just a few days out from the comsumer launch of AMD's next generation Vega graphics architecture, which will underpin its Radeon RX Vega family, and there is already talk about what's coming after Vega; and that is Navi. Not much is known about Navi architecture, other than vague descriptions citing "scalability" and "next generation memory" in addition to the chip being built on a 7nm FinFET process when it launches next year. Today, however, a new report is suggesting that Navi will be the first GPU from AMD with circuity that is dedicated to artificial intelligence (AI) operations. Apparently, AMD's Radeon Technologies Group wanted to enable such functionality in Vega, but just didn't have the time/resources to make that leap. NVIDIA was able to implement AI circuity into its Tesla V100 (Volta GV100) GPU that has 640 dedicated Tensor cores to tackle heavy AI workloads. AMD's Navi GPUs will be built using GlobalFoundries' 7nm node and dedicated Tensor cores could offer an order of magnitude processing lift for the company's next generation GPU.
This discussion was created for logged-in users only, but now has been archived. No new comments can be posted.

AMD Next Gen Navi GPU Architecture To Employ AI-Dedicated Circuitry

Comments Filter:
  • I wonder when AI-accelerating hardware is coming to smartphones. Google has developed dedicated chips. Apple is heavily into AI. You can imagine an on device AI chip offloading part of the processing from the cloud. In a way Apple's Homepod will already do that for Siri commands, and they are launching a Core Machine Learning library so I can imagine more powerful apps coming if helped by an AI savvy chipping the next generation phones, be it iPhones or Google Pixel phones. Interesting.

Recent investments will yield a slight profit.