Driven by AMD Instinct MI300, the strongest supercomputing El Capitan has begun to be assembled.

CTOnews.com, July 6 (Xinhua)-- Lawrence Livermore National Laboratory (Lawrence Livermore National Laboratory,LLNL) announced on Wednesday that it has begun assembling the El Capitan supercomputer and plans to launch it next year.

LLNL tweeted:

We have started to install El Capitan to implement the first 2 ExaFLOPS supercomputer.

We plan to launch in 2024, and there is still a long way to go, but it is really exciting that years of hard work will finally come true.

El Capitan and Frontier, Aurora and other supercomputing, HPE-based Shasta supercomputer architecture, also means that it is built by HP Enterprise.

The key components of El Capitan are AMD's upcoming Instinct MI300,CTOnews.com. As previously reported, Instinct MI300 APU will be manufactured in 5nm process, with 24 Zen 4-architecture CPU modules and CDNA 3-architecture GPU modules, with 128GB's HBM3, next-generation Infinity Cache, and fourth-generation Infinity architecture to support CXL 3.0 ecosystem.