New M1076 Analog Matrix Processor delivers performance, scalability and efficiency.
AI hardware company Mythic has released its latest AI processor, the M1076 Analog Matrix Processor (AMP).
As artificial intelligence applications (AI) continue to permeate many industries, the technology required to keep them functioning at the network edge has lagged in performance and power. Mythic has been working to break down those edge barriers and help keep up with rapid advances in AI with its analog compute-in-memory technology. The company believes its Analog Matrix Processor for AI applications represents a breakthrough in hardware development for AI.
Edge AI development has faced the challenge of dealing with large amounts of sensor data that need to be computed at extremely high speeds. Mythic has addressed this challenge through analog computations inside dense flash memory arrays. The result is a low-power, small-sized and cost-effective processor that enables the deployment of edge AI for multiple industries.
The latest iteration, the M1076 AMP, is a single chip that relies on the Mythic Analog Compute Engine to support up to 25 TOPS of AI compute. In a 3W power envelope, the M1076 AMP requires 10 times less power than typical system on a chip (SoC) or GPU solutions, according to Matrix.
In a 16-chip configuration, the M1076 AMP PCIe card can reach 400 TOPS of AI compute running on 75W. The M1076 AMP was built for scalability from the edge to servers and is designed for markets such as smart cities, industry, enterprise and consumer devices. The processor is suited for industrial machine vision, autonomous drones, surveillance cameras and network video recorders. It also supports cutting-edge augmented and virtual reality (AR/VR) applications with low latency human body pose estimations, which could play a large role in future smart fitness, gaming and collaborative robotics devices.
“Mythic’s groundbreaking inference solution takes AI processing and energy-efficiency to new heights,” said Tim Vehling, senior vice president of product & business development at Mythic. “With its ability to scale from a compact single-chip to a powerful 16-chip PCIe card solution, the M1076 AMP makes it easier for developers to integrate powerful AI applications in a wider range of edge devices that are constrained by size, power, and thermal management challenges.”
Despite the M1076 AMP’s performance and power, it is an ultracompact PCIe M.2 A+E Key card measuring just 22 mm x 30 mm. This bodes well for space-constrained embedded edge AI applications. In addition, for workloads with multiple streams or large, deep neural networks, a PCIe card form-factor with 16 Mythic AMPs can be used, supporting up to 400 TOPS and 1.28 billion weights.
In May, Mythic announced that it had raised $70 million in Series C funding, led by BlackRock and Hewlett Packard Enterprise, bringing its total funding to $165.2 million in total funding. The funding will go toward mass producing AI solutions to grow the company’s customer base.
“Customers are finding that powerful AI leads to significant differentiation of their products, and our cost-effective solution lets them deploy AI at an unprecedented scale,” said Mike Henry, CEO, cofounder and chairman of Mythic.
