RISC-V pioneer SiFive has announced a partnership with graphics and Arm-based processor giant Nvidia to integrate the former’s RISC-V core intellectual property with the latter’s deep-learning acceleration IP in a single high-performance system-on-chip (SoC) design.
“Nvidia open sourced its NVDLA [Nvidia Deep Learning Architecture] architecture to drive the adoption of AI,” explains Deepu Talla, vice president and general manager of autonomous Machines at Nvidia, of his company’s contribution to the project. “Our collaboration with SiFive enables customised AI [Artificial Intelligence] silicon solutions for emerging applications and markets where the combination of RISC-V and NVDLA will be very attractive.”
“It is great to see open-source collaborations, where leading technologies such as NVDLA can make the way for more custom silicon to enhance the applications that require inference engines and accelerators,” adds Yunsup Lee, co-founder and chief technical officer at SiFive. “This is exactly how companies can extend the reach of their platforms.”
The first demonstration of the partnership, which connects a field-programmable gate array (FPGA) running Nvidia’s NVDLA IP to a SiFive HiFive Unleashed development board’s Freedom U540 RISC-V processor via ChipLink, will take place at the Hot Chips conference this week, the companies have confirmed. Performance details have not yet been provided, though SiFive claims the resulting creation suitable for “intelligence at the edge, where high-performance with improved power and area profiles are crucial.”