What is world’s fastest super computer Fugaku.
This super computer has currently ranked as the most powerful in the world. It has progressed by RIKEN and Fujitsu starting back six years ago in 2014 . The development of fugaku was take it of becoming the core of Japan’s High Performance Computing Infrastructure.
The world’s most powerful supercomputer Fugaku is now fully developed in Japan. And the machine is available for research use.
The journey from tale to becoming top super computer.
- 23 May 2019 : RIKEN was introduced Fugaku as a supercomputer.
- August 2019 : . The logo of Faguku was launched, which is represent by Mount Fuji and illustrate as a “Fugaku’s high performance” and “the extensive range of its users”.
- November 2019 : The representative of Fugaku won first place in the Green500 list.
- 2 December 2019 : The RIKEN facility began the shipment of the equipment.
- 13 May 2020 : All shipment has to be completed.
- june 2020 : Fugaku became the fastest supercomputer in the world(IN TOP 500)list,which was display by the IBM summit.
However, Fugaku has currently ranked as the most powerful super computer in the world right now.
Fugaku has topped the Top500 list, a supercomputer benchmark index, for two consecutive years. The computer has 100 times the application performance of K supercomputer. And is develope to implement high-resolution, long-duration and large-scale simulations.
what is Exascale computing?
Exascale computing systems efficiently measure at least 1018 floating point operations per second. Moreover, the name generally refers to the performance of supercomputer systems and even though no single machine has reached this goal within January 2021. And there has a system being design to reach this milestone.
It is the first ever supercomputer which achieved 1 exaFLOPS. As of April 2021, Fugaku is becoming the fastest computer in the world.
Software use in Fugaku.
Fugaku has built with the “light-weight multi-kernel operating system” which is name as IHK/McKernel. However, Fugaku uses both Linux and McKernel light-weight operating system. The infrastructure that both kernels run on is termed the Interface for Heterogeneous Kernels (IHK). Additionally, it is available with the high-performance simulations McKernel. And on the other hand Linux available for all other POSIX-compatible services.
Besides the system software this super computer able to operate many kind of application,including several benchmark. However, operating the main stream HPL benchmark, Fugaku comes with the petascale and halfway to exascale. Additionally, Fugaku has create world records on at least three other benchmarks, including HPL-AI at 2.0 exaflops, the system has exceed the exascale threshold for the benchmark. A representation of that benchmark is come as following:
However, the troubleshooter method of choice has a mixture of LU factorization and repetition refinement performed afterwards to bring the solution back to 64-bit accuracy. The change OF HPL-AI lies in dropping the requirement of 64-bit computation throughout the entire solution process and instead opting for low-precision (likely 16-bit) accuracy for LU, and a sophisticated iteration to recover the accuracy lost in factorization.
Hardware use in Fugaku.
Fugaku has built with the Fujitsu A64FX microprocessor technology. In addition to this, Fugaku comes with ARMversion 8.2A CPU. Undoubtedly Fugaku became more power then its previous version the K computer .
However, the configuration of Fugaku used 158,976 A64FX CPUs. Which alternately come together with the using Fujitsu’s proprietary torus fusion interconnect. Moreover, it upgrade in November with the increase number of prosesser.
As a matter of performance, it comes with Rmax of 416 petaFLOPS in the FP64high performance LINPACK benchmark. A performance increased to a Rmax of 442 petaFLOPS after the upgrade.
Moreover with the upgrades of processor will also increase the other ranking that test computers on different workload. And it has also includes Graph 500, HPL-AI, and HPCG technology. However no other computer is able to do four rankings at a once.
Supercomputer is increasing it’s performances on the new different precision HPC-AI benchmark to 2.0 exaflops, testing its 1.4 exaflops mark recorded six months ago. These represent the first benchmark measurements above one exaflop for any precision on any type of hardware.” (a 42% increase).
Impressively, the Arm A64FX core-count was only increase by 4.5%, to 7,630,848, but the calculated performance rise much more on that benchmark (and the system does not use other compute capabilities, such as GPUs), and a little more on TOP500, or by 6.4%, to 442 petaflops, a new world record and widening the opening to the next computer by that much. For the High-Performance Conjugate Gradient (HPCG) benchmark it’s over 5.4x as fast, at 16.0 HPCG-petaflops, as the number two system, Summit, that happens to also be second on TOP500.
Fugaku’s production outshine the unite performance of the next 4 supercomputers on the top500 list (almost next 5) and outclass by 45% margin all the other top-10 computers on HPCG benchmark.