The King Abdullah University of Science and Technology (KAUST) in Saudi Arabia has announced that HPE has won the bid to build the Shaheen III supercomputer. Shaheen III, slated for full operation next year, is expected to deliver 100 Linpack petaflops from its accelerated partition alone, easily making it the most powerful supercomputer in the Middle East.

Shaheen III will consist of a compute-only partition and an accelerated partition. The compute-only partition, spread across 18 liquid-cooled HPE Cray EX4000 cabinets and networked with HPE Slingshot, will comprise 4,608 nodes, each with two of AMD’s fourth-gen Epyc “Genoa” CPUs. KAUST is not yet disclosing performance estimates for the CPU cabinets at the moment — perhaps because providing estimates would allow one to infer the performance of the as-yet undetailed Genoa CPU. As a reference point, though, if Genoa matched the 3.58 FP64 teraflops of the AMD’s current-gen, top-bin Milan CPU, the 9,216 AMD CPUs in Shaheen III would deliver around 33 peak petaflops.

Shaheen III will also boast seven accelerated cabinets (for a total of 25 EX4000 cabinets). Across these seven accelerated cabinets: 704 nodes, each with quadruple Nvidia “Grace Hopper” Superchips, each with a tightly coupled CPU and GPU. These 2,816 Superchips will contribute the bulk of Shaheen III’s flops: HPE says that this partition alone will weigh in at 100 Linpack petaflops. While Nvidia has yet to release teraflops for the Superchips, we might reasonably speculate from this information that each Grace Hopper Superchip might deliver somewhere in the range of 35 Linpack teraflops.

Based on the accelerated partition alone, Shaheen III will be around 20 times faster than Shaheen II, which debuted in 2015 with 5.54 Linpack/7.24 peak petaflops (and which itself was 25 times faster than its predecessor). This formidable system will be accompanied by an expansion to Shaheen II’s storage: 50PB of HPE’s Cray ClusterStor E1000 storage.

“A supercomputer like Shaheen III is a universal scientific instrument employed by scientists and engineers in every discipline for tasks such as simulation, analysis of experimental data, learning from observed data, and efficient data storage and retrieval,” said David Keyes, director of KAUST’s Extreme Research Computing Center (ECRC) and a professor of applied mathematics and computational science. “It is the ultimate scientific ‘watering hole’ at which researchers of different disciplines exchange techniques and software tools. An advance in one field spurs advances in several.”

(KAUST and Keyes just recently landed HPCwire coverage due to their involvement in “Reshaping Geostatistical Modeling and Prediction for Extreme-Scale Environmental Applications,” one of the six projects nominated for this year’s Gordon Bell Prize. The project used both Shaheen II and Fugaku.)

KAUST will use Shaheen III to tackle topics like clean combustion, ecosystems in the Red Sea, climate modeling, solar energy, preventative healthcare, drug discovery, crop resilience and the Arabian tectonic plate. More broadly, the supercomputer will enable KAUST to scale up its simulations, analyses and — in particular — its AI work, which will benefit from HPE’s Machine Learning Development Environment. The partners framed the investment in the new supercomputer in terms of Vision 2030, a strategic plan aimed at diversifying the country’s economy.

“The new HPE Cray EX system will allow us to conduct research on a larger scale, resulting in significant scientific, economic and social advances,” said Tony Chan, president of KAUST. “In line with Vision 2030, we strive to meet the ever-increasing demands of our active and solutions-driven faculty, and also those of external partners, for faster and more efficient computing resources. KAUST supercomputing resources are used by more than half of our faculty, students, postdoctoral students and researchers, and researchers from more than 20 external organizations in the Kingdom.”

Absent emerging challengers, Shaheen III will become the most powerful (disclosed) supercomputer in the Middle East, taking over from Saudi Aramco’s 22.4 Linpack petaflops Dammam 7 supercomputer, which is also housed in Saudi Arabia and which was also built by HPE.