Inspire Semiconductor announced the development of its second-generation ASIC solution for Artificial Intelligence (AI), Blockchain and Scientific High Performance Computing.
InspireSemi spun out of CryptoCore Intellectual Holdings, which also provided seed funding, and is dedicated to delivering superior computation solutions for AI and other compute intensive industries. Its Gen1 ASIC chip is already a powerful performer in the crypto mining space.
“We couldn’t be more pleased with the results of our Gen1 solution,” said InspireSemi Founder and President, Alex Gray. “Gen2 will build on Gen1 with a more versatile architecture that integrates with established software ecosystems and sets new standards in performance and power efficiency.”
InspireSemi’s versatile processor architecture is based on an array of thousands of simple but powerful CPU cores, tightly integrated with memory and networking. Evolved for ultra-efficient blockchain applications, it is also an ideal platform for high-performance machine learning and scientific computing.
The highly optimized, proprietary on-chip network removes crucial bottlenecks from applications that depend on close cooperation between many processors. Seamless arrays of hundreds of chips can be constructed, approaching one million cores.
Compared to a modern datacenter GPU, the InspireSemi system provides all the raw computational power with greater flexibility and agility. The advanced network fabric enables better cooperation between cores with low latency and extreme bandwidth. The company’s roots in power and cost-sensitive blockchain computing enable InspireSemi to provide this performance with greater efficiency than leading GPUs.
InspireSemi delivers a turn-key compute system including its proprietary power delivery and liquid cooling solutions.
“Gen2, by design, will be capable of providing solutions for a far broader range of verticals than Gen1,” said InspireSemi Chief Financial Officer, John B. Kennedy. “Currently, we’re focusing on dynamic markets that are in need of extreme compute power, energy efficiency and affordability.”
