Supercomputer-on-a-chip goes live: one PCIe card contains more than 6,000 RISC-V cores, with the ability to scale to more than 360,000 cores – but the startup still remains elusive on price
InspireSemi has announced the successful tapeout of the Thunderbird I Accelerated Computing chip for manufacturing at TSMC.
This highly differentiated “supercomputer cluster-on-a-chip” features 1,536 custom 64-bit RISC-V CPU cores, tailored for high-level scientific computing and complex data processing.
Thunderbird I is designed to meet a wide range of compute-intensive applications, from AI and machine learning to graph analytics. Leveraging the open standard RISC-V CPU ISA, it enables easier development and integration into existing technology frameworks, with access to a robust ecosystem of software, libraries and tools.
PCIe expansion card planned
The chip’s architecture integrates a high-speed mesh network fabric that provides significant bandwidth and minimal latency communications between cores, important for applications that rely on synchronized operations across multiple threads. This efficient network integration manages the interactions within the chip’s core array and memory systems, ensuring optimal performance without the usual bottlenecks.
The upcoming product release includes a server PCIe expansion card with four Thunderbird chips, offering more than 6,000 interconnected 64-bit CPU cores. This setup is equipped for double-precision mathematics, essential for many high-performance computing applications in areas such as climate science, medical research and complex simulations.
Ron Van Dell, CEO of InspireSemi, said: “We are proud of the achievement of our engineering and operations team to complete the Thunderbird I design and submit it for production to our world-class supply chain partners, TSMC, ASE and imec. We expect to start deliveries to customers in the fourth quarter.” However, there’s no word yet on the price.
InspireSemi also highlights Thunderbird I’s energy efficiency, a holdover from its original design for energy-sensitive blockchain computing applications. The company says this approach provides a more environmentally friendly alternative to traditional data center GPUs.