NVIDIA A100 LIQUID COOLED PCIe | Meeting Customer Demands for High-Performance, Green Data Centers

The NVIDIA A100 Liquid Cooled Tensor Core GPU for PCIe delivers the performance required while using far less power-at every scale-to propel the world's highest performing elastic data centers for AI, data analytics, and high performance computing (HPC) applications. As the engine of the NVIDIA data center platform, A100 provides up to 20x higher performance over the prior NVIDIA Volta generation. A100 can efficiently scale up or be partitioned into seven isolated GPU instances, with Multi-Instance GPU (MIG), providing a unified platform that enables elastic data centers to dynamically adjust to shifting workload demands, while offering a road to sustainability as well.

Switching all the CPU-only servers running AI and HPC worldwide to GPU-accelerated systems would save a massive 11 trillion watt-hours of energy a year, the energy consumed by more than 1.5 million homes in a year. Liquid cooling saves water and power by eliminating chillers that evaporate millions of gallons of water a year to cool the air inside data centers, by delivering systems that recycle small amounts of fluids in closed systems focused on key hot spots. NVIDIA estimates that liquid cooled data centers could hit 1.15 PUE (Power Usage Effectiveness), far below the 1.6 for air cooled counterparts. Liquid cooled data centers can pack twice as much computing performance into the same space, too. That's because the NVIDIA A100 Liquid Cooled GPU only uses one PCIe slot, while the passively air cooled A100 GPU board requires two-resulting in up to 66% fewer racks required, and 28% lower power consumption for each rack.