RISC-V Tensor Unit Enables Energy Efficient AI Processing
27 October 2023
Intended for use in ultra-fast artificial intelligence (AI) implementations, a RISC-V Tensor Unit has just been introduced by Semidynamics.
Relying on the company’s configurable 64-bit cores, this presents customers with hardware that is highly optimised to meet the requirements of matrix multiplication workloads. The result is a solution that can handle intense computation demands while still keeping the power budget involved to a minimum.
The unit is built on top of the Semidynamics RVV1.0 vector processing resource, and utilises the existing vector registers to store matrices. Since it does not require a new architecturally-visible state, it is applicable to any RISC-V vector-enabled Linux without the need for any adaptation.
According to Semidynamics’ CEO, Roger Espasa; “This new Tensor Unit is designed to fully integrate with our other innovative technologies to provide solutions with outstanding AI performance. First, at the heart, is our 64-bit fully customisable RISC-V core. Then our Vector Unit which is constantly fed data by our Gazzillion technology so that there are no data misses. And then the Tensor Unit that does the matrix multiplications required by AI. Every stage of this solution has been designed to be fully integrated with the others for optimal AI performance and very easy programming. The result is a performance increase of 128x compared to just running the AI software on the scalar core.”
Contact Details and Archive...