Achronix Semiconductor Corporation: Pushing Boundaries in High-performance Compute Market

Application Specific Integrated Circuits (ASICs) are increasingly being used as processing elements in high-performance compute applications such as AI/ ML, cryptocurrency and 5G wireless infrastructure that, until recently, were reserved for traditional CPUs. A typical setup here is a dual-processing system with the ASICs acting as an accelerating co-processor offloading and enhancing the performance of the CPU. However, key challenges in leveraging ASICs for HPC workloads are lack of adaptability and programmability on-chip. With algorithms evolving or changing regularly for high-performance compute applications, taping out a new chip is the only option to adapt the ASIC, which is expensive (costs running in tens of millions of dollars) and takes many months to complete. This results in slower time-to-market and in turn loss of market share due to increased competition from incumbents or new players entering the space.

Achronix Semiconductor Corporation, a California-based firm, addresses this challenge and helps businesses push the boundaries of the high-performance compute market by providing embedded Field Programmable Gate Array (eFPGA) solutions. With Speedcore™ eFPGAs, a new or more efficient algorithm can be easily downloaded and updated in the ASIC as required, avoiding the need to make a new chip. Historically, companies have either used standalone FPGAs (for flexibility, programmability) or ASICs (for high-performance, high-volume applications). But Achronix’s Speedcore eFPGA defines a new paradigm in data acceleration for HPC applications by combining the flexibility and programmability of FPGAs with the performance and specialized application implementation of ASICs.

“High-performance compute applications require new and innovative approaches, and Achronix is leading the HPC market with the Speedcore eFPGA solutions,” remarks Steve Mensor, Vice President of Marketing, Achronix Semiconductor Corporation.