The US dept of Energy has inked a deal for a new supercomputer. The contactors are Intel and Hewlett Packard Enterprise (HPE). Recall that HPE bought Cray who used to be the big vendor of supercomputers. Cray was a very well known company.
Powered by Intel’s latest technologies, Aurora will be fueled by a future generation of Intel Xeon Scalable processor (code-named “Sapphire Rapids”) on enhanced SuperFin technology, Intel Xe HPC GPU (“Ponte Vecchio”), a future generation of Intel Optane persistent memory and Intel oneAPI software.
The new machine is designed for over 1 EFLOPS of performance. For comparison the RTX 2080 is bit over 10 TFLOPS. The AMD Ryzen R5 3600 is about 1,209.6 GFLOPS when using SSE and AVX capabilities.
Can Aurora play Halo? Most likely it can handle even 16360×8640 (16K) without a problem. Some use Crysis for game performance but the remastered Halo is 64-bit which makes it easier to leverage modern hardware.
The Fugaku is the current leader for supercomputers (415.5 PFLOPS) but when Aurora is finished next year, it will be on top only until somebody builds an even more powerful machine.
Scientists will use Aurora to build a software environment that will bring the combined power of exascale computing and deep learning to advance cancer research, diagnosis and treatment.
Aurora will help scientists better understand the structure of the brain and how each neuron connects with others to form the brain’s cognitive pathways. This research could lead to potential cures for neurological diseases.
Researchers will use Aurora to advance the design of more efficient aircraft by enabling advanced computational modeling capabilities that provide deeper insights into how turbulent flows affect fuel consumption and performance.
Using Aurora, scientists will advance fusion energy research by leveraging artificial intelligence to predict and tame conditions that can halt fusion reactions and damage tokamak devices.
Aurora will advance computational catalysis, providing tools and capabilities to help scientists develop novel catalysts that can improve fuel production and other important industrial processes.
Researchers will use Aurora to develop exascale workflows and algorithms and machine learning capabilities to accelerate the search for new physics discoveries at CERN’s Large Hadron Collider. CERN has long needed extreme supercomputers to help analyse the results of experiments.
An earlier DoE supercomputer was used recently to study COVID-19 which make vaccine development substantially faster. China published the genome of the coronavirus and it was then analysed all over the world.