The RISC-V Summit, held recently in Santa Clara, California, featured presentations from major players in the technology industry, including Nvidia, Qualcomm, Google and Samsung. Their participation underscored the growing role of open processor architecture, which is increasingly seen as a competitor to proprietary Arm and x86.
Image source: Pete Linforth / pixabay.com
Nvidia has been using RISC-V solutions in GPU microcontrollers for almost a decade. The company presented a 20-minute presentation at the summit entitled “One Architecture, Dozens of Applications, Billions of Processors,” delivered by Frans Sijstermans, Vice President of Multimedia and ASIC Architecture. He discussed how the company is using RISC-V to improve its products, and emphasized that open architecture is increasingly influencing GPU design.
Qualcomm, which previously announced a long-term commitment to RISC-V, has raised the possibility of using Sail and other alternative approaches to develop a common RISC-V instruction set architecture (ISA) standard. The company, along with Nvidia and other colleagues, also took part in a section dedicated to the prospects of AI and security issues.
Samsung shared its experience of successfully introducing RISC-V processors into embedded systems and explained how its semiconductor manufacturing division is helping customers deploy innovation using open processor architecture. The company discussed ways to optimize the performance of chips and chiplets.
The presentation by Cliff Young and Martin Maas from Google DeepMind was one of the highlights of the event. They talked about the prospects for creating new AI accelerators with RISC-V. Google TPU chips are based on this architecture – the company talked about their successful deployment and the problems that arose.
The widespread introduction of RISC-V into server systems and PCs will take a long time, but the architecture is already gaining momentum in the field of AI and the automotive industry. Other speakers discussed how it could set new trends in computing and the role RISC-V could play in generative AI and HPC.