Ingrasys Highlights AI Supercomputing Power at SC24
Atlanta, Georgia (Nov. 19, 2024) – Ingrasys, a global leader in cloud infrastructure solutions, recently attended Supercomputing 2024 (SC24), an international high-performance computing conference, being held at the Georgia World Congress Center from November 17-22. Visitors can experience Ingrasys's latest innovations at booth #2845, where the company will demonstrate its state-of-the-art accelerated data center infrastructure solutions that are driving the AI revolution forward.
Propelling AI innovations forward, Ingrasys will showcase its innovative computing solutions based on the NVIDIA Blackwell architecture. The NVIDIA GB200 NVL72 platform, the breakthrough liquid-cooled AI rack, interconnects 72 NVIDIA Blackwell GPUs and 36 NVIDIA Grace CPUs within a 72-GPU NVLink domain to function as one giant accelerator, which is tailored to trillion-parameter large language model (LLM) training and real-time inference. To meet diverse data center requirements, Ingrasys offers both liquid-to-air and liquid-to-liquid cooling options to support NVIDIA GB200 NVL72, achieving optimal data center performance and energy efficiency.
The exhibition will also spotlight the NVIDIA MGX™ systems to address any accelerated computing needs of modern data centers. Purpose-built to deliver unparalleled performance for mainstream LLM inference, the 4U air-cooled server supports up to 8 NVIDIA H200 NVL GPUs. This solution is ideal for enterprises with data center space constraints, unleashing AI acceleration for diverse AI and HPC workloads at any scale.
"We are committed to pushing the boundaries of AI computing capabilities while maintaining operational efficiency and sustainability," said Ingrasys. "Our advanced cooling solutions demonstrate our dedication to providing cutting-edge infrastructure. These innovations not only optimize AI performance but also significantly reduce data center Power Usage Effectiveness (PUE), enabling our customers to harness the full potential of AI technology while minimizing environmental impact."