Intel & Argonne National Laboratory announces the successful blade installation in the Aurora supercomputer, bringing it one step closer to full functionality.
The Intel-based Aurora supercomputer has 2 exaFLOPS of processing power and may surpass AMD’s limits
The Aurora supercomputer has endured several delays since its inception, but we may finally see it live. For those who don’t know, the Aurora supercomputer features Intel’s Xeon CPU Max and Xeon GPU Max lineup, boosting its performance to 2 ExaFLOPS. One of the applications of the Aurora platform will be to provide a state-of-the-art generative AI model for science.
It offers 10,624 nodes with 21,248 Xeon CPUs from the Sapphire Rapid SP series. It is equipped with a total of 63,744 GPUs based on the Ponte Vecchio design, offering a peak injection of 2.12 PB/s and a peak halving bandwidth of 0.69 PB/s.
Here’s the benefit of the Intel-based Aurora supercomputer, as previously detailed by Jeff McVeigh, Vice President of the Intel Super Compute Group:
- The Intel Data Center GPU Max Series outperforms the Nvidia H100 PCIe card by an average of 30%1 across various workloads, while independent software vendor Ansys shows a 50% acceleration of the Max Series GPU over the H100 in AI-accelerated HPC applications.
- The Xeon Max-series CPU, the only x86 processor with high-bandwidth memory, shows a 65% improvement over AMD’s Genoa processor in the High Performance Conjugate Gradients (HPCG) benchmark1 and consumes less power. High memory bandwidth is recognized as one of the most desired features for HPC customers.
- The scalable processors Intel .
- The Gaudi2 deep learning accelerator offers competitive performance in deep learning training and inference with up to 2.4x faster performance than Nvidia A100.
In terms of storage capacity, the Aurora supercomputer features 10.9PB of DDR5 system DRAM, 1.36PB of HBM capacity from the CPUs, and 8.16PB of HBM capacity from the GPUs. In addition, it uses an array of 1,024 storage nodes with a total capacity of 220 TB. If you are curious about how this gargantuan system is used, below is a brief explanation:
From fighting climate change to finding cures for deadly diseases, researchers face daunting challenges that require advanced computing technologies at scale. Aurora is ready to listen to the needs of the HPC and AI communities and provide the necessary tools to push the boundaries of scientific research.
The latest Intel Data Center GPU Max Series 1550 running on Aurora offers the best SimpleFOMP performance, outperforming the NVIDIA A100 and AMD Instinct MI250X accelerators. However, the supercomputer has yet to pass preliminary tests. After that, it is expected to appear in the Top500.org list, potentially overtaking the AMD-based Frontier supercomputer. The Aurora supercomputer is expected to be fully operational later this year.