The FTSAI Lab is outfitted with cutting-edge technology designed to facilitate advanced research in FinTech and AI while ensuring durability and long-term efficiency. At the heart of its infrastructure are three highly adaptable supercomputers equipped with scalable processors that feature built-in acceleration technology and support the latest platform innovations.
These systems are powered by two fourth-generation Intel Xeon processors (Sapphire Rapids), optimized for AI and ML workloads with integrated accelerators. They support state-of-the-art technologies such as DDR5 RAM, PCIe Gen5 for high-speed peripheral communication, and a higher core count than previous Xeon generations. The motherboard is also designed to accommodate DDR5 RAM with a maximum frequency of 4800MHz and PCIe Gen5 buses, doubling the speed of its Gen4 predecessor.
Furthermore, each system is equipped with an Nvidia A100 GPU, a top-tier choice for AI and ML applications. Based on Nvidia’s Ampere architecture, this GPU boasts 6912 CUDA cores, delivers 19.5 teraflops of FP32 performance, and includes 80GB of memory. The A100 supports Multi-Instance GPU (MIG) technology, enabling it to be divided into seven independent partitions, allowing multiple models to be trained simultaneously without compromising data integrity or performance.
With its state-of-the-art capabilities, this AI computing unit delivers exceptional performance, scalability, and versatility, making it a powerful asset for AI research, data analytics, and high-performance computing.