Esperanto Technologies has announced that it can run generative artificial intelligence models on low-power RISC-V hardware systems. The company intends to offer generative AI technology models to researchers working within the RISC-V community to accelerate development.
Esperanto Technologies develops large language models on low-power hardware systems to reduce the total cost of ownership. The company thinks utilizing generative AI models on RISC-V platforms will encourage research and development in AI and general-purpose applications from the cloud to the edge.
“Generative AI is one of the latest advancements in machine learning, and we are pleased to contribute elements of our efforts in the area of large language models to the RISC-V research community,” says Art Swift, the president and CEO at Esperanto Technologies.
Esperanto Technologies has successfully implemented several versions of Meta’s Open Pre-Trained Transformer (OPT) model on its hardware. The company showed that its chips have a low power consumption of 25W per chip for inferencing. They further demonstrated the flexibility and efficiency of its platform by porting OPT models onto its ET-SoC-1 silicon chipset.
The ET-SoC-1 chip uses 24 billion transistors based on the 64-bit RISC-V instruction set architecture for performance-demanding machine learning applications. The company employs 7nm process technology to increase the chip’s processing power. There are 1088 ET-Minion in-order processor cores with vector and tensor units as inference accelerators. Additionally, the hardware has ET-Maxion out-of-order processor cores to support the operating system.
“We are excited to be working with Esperanto to extend the deployment of its RISC-V solutions to a broader set of customers that are searching for low-power AI inference solutions and a reduced total cost of ownership,” says Thierry Pellegrino, the president of Penguin Solutions and SVP of intelligent platform solutions at Smart Global Holdings.
There have been other recent developments regarding the process of porting generative AI models onto embedded systems solutions. For example, Siemens and Microsoft have teamed up to use generative AI models to speed up the development of factory automation.
AI/ML | Esperanto Technologies | generative AI | IoT | Microsoft | model | RISC-V