Exxact Corporation, a leading provider of HPC, artificial intelligence (AI), and data center solutions, has announced that it now offers a line of TENSOREX™ servers featuring NVIDIA HGX™ A100. In order to maximize scientific progress, researchers and scientists can combine simulation, data analytics, and artificial intelligence with this new line of GPU-accelerated systems.
Andrew Nelson, VP of Technology at Exxact Corporation, said, “Compared to previous generations, HGX provides up to a 20X AI speedup out of the box with TensorFloat 32 (TF32) and a 2.5X HPC application speedup with FP64 tensor cores. NVIDIA HGX delivers a staggering 10 petaFLOPS, forming the world’s most powerful accelerated scale-up server platform for AI, data analytics, and HPC. The NVIDIA HGX platform has revolutionized a broad range of industries with over 2,000 applications optimized for GPUs to help you accelerate your work. NGC offers GPU-optimized and tested application containers for all the major deep learning frameworks to ensure that you’re getting the most from your GPU infrastructure.”
He added, “As compute for AI models become more demanding, the NVIDIA HGX A100 with NVIDIA NVSwitch has the performance to enable training of massive neural networks with its unified memory architecture.”
It combines NVIDIA A100 Tensor Core GPUs with high-speed interconnects to create servers that are superior to those encountered in other places. In addition to it having 16 A100 GPUs, HGX has 1.3 terabytes (TB) of GPU memory and memory bandwidth of more than 2 terabytes per second (TB/s) for an unprecedented computing experience.