Supercomputers have been constantly upgraded and updated to match the requirements of technological advancements. Leonardo 4th most powerful supercomputer has been included in Top500.USA’s Frontier is the most powerful supercomputer. The USA has 50 % of the world’s total computing power. China is currently leading the list with 173 supercomputers, with the USA in second place. Leonardo will serve a range of the most demanding applications, including materials science, biomedicine, climate change, engineering, modeling of the human brain, and AI development.
The LUMI system in Finland was the first one to enter into the service in June 2022. On November 24th, Italy inaugurated Leonardo, the fourth most powerful supercomputer, at the Bologna Technopole.This means Europe will now have 2 of the most powerful supercomputers. The other two are US Frontier exascale system and Japan’s Arm-based Fugaku system. The top four ranked supercomputers in the Top500 are:
- US Frontier system
- Japan’s Fugaku
- EuroHPC’s LUMI
- Leonardo pre-exascale system
Supercomputer vs Mainframe computer
Seymour Cray invented Supercomputers.. They are larger in size with high speed and are expensive. Their primary function is to help in complicated mathematical operations. They can execute billions of instructions in just a second.
OpenAI to train bots for complicated tasks
IBM invented Mainframe computers. They are advanced, smaller in size, and slow in speed which can store large amounts of data. They are economical and serve the basic purpose effectively.
Leonardo’s specifications
French IT outfit Atos made Leonardo, the 4th most powerful supercomputer, which is based on its BullSequana XH2000 architecture. It includes two main computing modules, Booster and Data-Centric. These computing modules will help to cover a wide range of workloads.
Capable of delivering over 249 petaflops in performance.
The system is equipped with 3,500 Intel Xeon processors and 14,000 Nvidia A100 GPUs. The Booster comprises 3456 Intel Ice Lake computing nodes, and the Data Centric module has 1536 nodes. It will be using quantum processors as accelerators to be more efficient.
Frontier supercomputer remains #1
Frontier remains unbeaten and remains #1 in the TOP500. It is the first US system that has exceeded one ExaFlop/s. The US Department of Energy will operate it. ORNL will integrate and test it in Tennessee, US. The new HPE Cray EX architecture combines 3rd Gen AMD EPYC™ CPUs optimized for HPC and AI with AMD Instinct™ 250X accelerators and Slingshot-11 interconnect.
Green500: New Energy-Efficiency King
Lenovo for the Flatiron Institute in New York has developed it and is the only system to use Nvidia’s Hopper GPU architecture to make it Supercomputing leaderboards. It is similar to LUMI or Frontier, based on Lenovo’s ThinkSystem SR670 V2 server platform. Each node pairs two 32-core Ice Lake Xeon Scalable processors with four of Nvidia’s 80GB H100 GPUs. Henri is the second-smallest system on the list, with a total of 5920 cores between the CPU and GPUs
Andromeda supercomputer with more cores
The Andromeda supercomputer consists of 16 CS-2 systems. Each one of them is powered by Cerebras’s massive Wafer-Scale Engine 2 (WSE-2) chip and connected by the startup’s SwarmX interconnect fabric. With 850,000 compute cores, the chip contributes to Andromeda reaching the 13.5 million core mark, surpassing the 8.7 million AMD CPU and GPU cores of the Frontier.
‘AI supercomputer’ to use Nvidia GPUs and software
Microsoft and Nvidia are going to pair up to build an AI supercomputer. They will use Azure infrastructure supported by Nvidia’s GPU accelerators, network kit, and its software stack. Microsoft is the first public cloud that has incorporated its Quantum-2 InfiniBand networking switches for its AI-optimized virtual machines. The current Azure comprises 200Gb/s Quantum InfiniBand and A100 GPUs. 400Gbps Quantum-2 InfiniBand and the H100 GPUs will replace them in the future.