Original Coverage & Source Attribution: wccftech.com
Researchers in China have used exascale computers to solve a problem that was thought to be only solvable by quantum computers. The development was reported at the SC26 conference for high-performance computing earlier this month in Chicago, with researchers from seven universities, laboratories and research centers participating in the effort.
The problem solved by the researchers was to discover the lowest energies and quantum states of the metal-sulfur-iron core at the heart of the enzyme used to convert atmospheric nitrogen into ammonia.
Chinese Researchers Used 1,280 Nodes Of The LineShine Exascale Supercomputer To Solve Problem Used As Quantum Computing Benchmark
The problem that the Chinese researchers worked on involved calculating the lowest possible energy, or the ground state, of the chemical center of the enzyme used to convert atmospheric nitrogen into ammonia.
The enzyme’s chemical center, called the iron–molybdenum cofactor (FeMoco), has 113 electrons spread across 76 orbitals, which means that tracking all possible ways that electrons can arrange themselves across the orbitals is computationally intense.
Consequently, since the possible arrangements were too many to brute-force the problem and algorithms for estimates ended up being inaccurate, computing FeMoco’s ground state energy ended up being treated as a benchmark for future quantum computing performance.
However, the researchers in China were able to solve the problem of determining the ground state energy by using exascale supercomputers instead. According to the details, code for the problem benchmarked on 20,480 nodes of the LineShine exascale supercomputer and China’s New-generation Intelligent Supercomputer (CNIS) achieved 1.456 EFLOP/s (exaFLOPS per second) in double precision and 5.626 EFLOP/s in mixed precision in computation speed.
Yet, while the code was run on thousands of nodes, in the end, the researchers were able to solve the problem on 1,280 nodes in 105 minutes. The time expended for the solution is equally important as the breakthrough since the FeMoco benchmark for quantum computing performance assumes that a fault-tolerant quantum computer will solve it in 8.5 hours.
The 113-electron 76-orbital FeMoco model was defined as a benchmark in 2019, and earlier this year, other researchers had demonstrated that classical computing techniques could be used to calculate FeMoco’s ground state energy to cast doubt on whether classical computers were completely locked out of the benchmark.
About the author: Ramish is a seasoned technology writer and editor with more than a decade of experience. He specializes in semiconductor fabrication and market analysis.
With a background in finance and supply chain management – via his bachelor’s in Finance and a micromasters in supply chain management from MIT – Ramish can cover deep industry insight to deliver accurate and authoritative coverage.
Follow Wccftech on Google to get more of our news coverage in your feeds.
