Mythic has announced what is claimed to be the industry’s first Analog Matrix Processor (Mythic AMP).
The launch of the M1108 AMP indicates a compelling new era for AI, providing for the first time an analog compute solution that fulfils best-in-class performance and power with accuracy when compared to digital devices. The solution opens up outstanding possibilities for device deployment at the edge in a wide range of applications and markets, comprising smart home, AR/VR, video surveillance, drones, smart city, and automation on the factory floor.
Running AI models at the edge rather than in the cloud delivers significant advantages to manufacturers: designs can be simpler, privacy is greatly enhanced, and the overall user experience is outstanding. However, AI at the edge also indicates the deployment of millions of devices, which makes another set of challenges, such as the requirement for low-cost, small form-factor devices with low latency, high performance, and low power. Digital inference solutions based on SOCs, TPUs, CPUs, GPUs, and FPGAs can fulfil some of these challenges. Still, inherent limits due to memory, clock speeds, and process technology make insurmountable tradeoffs, with the outcome that only processors capable of high-performance analog compute, such as this solution, can address them all.
“This is a significant inflection point in the industry. We are delivering technology that was previously thought to be impossible,” said Mike Henry, co-founder, CEO, and chairman of Mythic. “Our Analog Compute Engine eliminates the memory bottlenecks that plague digital solutions by efficiently performing matrix multiplication directly inside the flash memory array itself. The high performance and low power of the Mythic AMP combine to open up AI technology to broader application areas and address product categories that are currently inaccessible to comparable digital solutions.”
The M1108 AMP will be offered in both PCIe M.2 and PCIe card form-factors, and M1108 PCIe evaluation kits are available on request.