FUJITSU Processor A64FX
World's Top-level Processor Featured in the Supercomputer Fugaku
Supercomputer Processor Accelerating Digital Transformation
Supercomputers are expected to play an increasingly major role in people’s lives, using simulations and data analyses to assist in solving various societal issues and spearheading the march toward the digital transformation of society. The A64FX processor developed by Fujitsu becomes the heart of supercomputers that can perform quick simulations and analyze large data sets. Supercomputers with this processor perform at a high level, are highly reliable and offer a strong performance vs. power ratio.
The A64FX is the world's first processor to implement Scalable Vector Extension (SVE), an extension of the Armv8.2-A instruction set architecture for supercomputers. It has 48 calculation cores and two or four assistant cores, and has a theoretical peak performance of 3.3792 teraflops in double precision floating point calculations. In addition, it can perform single-precision and half-precision floating-point calculations, as well as 8-bit and 16-bit calculations with 512-bit wide SIMD, all with high throughput, leading to strong results in HPC/AI and other applications.
Both memory performance and calculation performance are important in HPC systems.
The A64FX offers both, with 48 calculation cores, high memory bandwidth, 512bit wide SIMD and other cutting edge technology, resulting in excellent calculation performance in real-world applications where strong performance in both areas is needed.
Evolved Power Efficiency
HPC systems require technology with excellent power efficiency.
The A64FX features the latest 7nm process, 2.5D packaging technology and microarchitecture which produces optimal power consumption from applications. In real-world applications, this offers a strong power to performance ratio.
Extensive Data Integrity
Reliability is a must for large-scale parallel processing systems. The A64FX has approximately 128,400 error check circuits, allowing it to offer mainframe-class RAS.
HPC Platform Featuring A64FX Processor
RIKEN/Fujitsu Joint Development
Supercomputer Fugaku is the world's highest level supercomputer, and is able to handle a wide range of application software with a high degree of execution performance. RIKEN and Fujitsu are developing it jointly with the goal of shared use in fiscal 2021.
- Read More
FUJITSU Supercomputer PRIMEHPC FX1000/FX700
These supercomputers are based on technology from the supercomputer Fugaku and offer high performance, high scalability and high reliability with excellent power conservation.
- Read More
November 20, 2020Press Release
- Japanese Research Group Performs Largest Ever Meteorological Calculation on Supercomputer Fugaku
November 19, 2020Press Release
- Fujitsu, AIST, and RIKEN Achieve Unparalleled Speed on the MLPerf HPC Machine Learning Processing Benchmark Leveraging Leading Japanese Supercomputer Systems
November 17, 2020Press Release
- Japan's Fugaku Retains Title as World's Fastest Supercomputer