New solver simulates nanoelectronics at the atomic-level up to 10,000 atoms

Technology News |
By eeNews Europe

The research group led by Mathieu Luisier from the Integrated Systems Laboratory (IIS) at ETH Zurich started with a so-called quantum transport simulator software named OMEN, which runs its calculations based on what is known as density functional theory, allowing a realistic simulation of transistors in atomic resolution and at the quantum mechanical level. This simulation visualizes how electrical current flows through the nanotransistor and how the electrons interact with crystal vibrations, thus enabling researchers to precisely identify locations where heat is produced. In turn, OMEN also provides useful clues as to where there is room for improvement.

But so far, conventional programming methods and supercomputers only permitted the researchers to simulate heat dissipation in transistors consisting of around 1,000 atoms. Producing a realistic simulation of larger objects was impeded by a data communication bottleneck between the processors and memory requirements. Indeed most computer programs spend more time moving data between processors, main memory and external interfaces than performing actual computing operations. According to the scientists, OMEN also suffered from a pronounced bottleneck in communication, which curtailed performance.

Self-heating in a so-called Fin field-effect transistor (FinFET)
at high current densities. Each constituting Silicon atom is
coloured according to its temperature.
Credit: Jean Favre, CSCS

Until now, the parallelization of OMEN was designed according to the physics of the electro-thermal problem, as Luisier explains. Rather than look at the physics, the researchers now focused on the dependencies between the data so they were able to reorganize the computing operations according to these dependencies to optimize the code.

Using the best supercomputers available today, the resulting code dubbed DaCe OMEN produced simulation results that were just as precise as those from the original OMEN software, though up to 140 times faster.

In a paper titled “A data-centric approach to extreme-scale ab initio dissipative quantum transport simulations” published in the 2019 Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, the researchers reported a realistic simulation of transistors ten times the size, made up of 10,000 atoms, on the same number of processors and up to 14 times faster than the original method took for 1,000 atoms. This extreme boost in computing speed has earned the researchers the Gordon Bell Prize.

The scientists achieved this optimization by applying the principles of data-centric parallel programming (DAPP), which was developed by professor Torsten Hoefler’s research group at ETH Zurich.

ETH Zurich –



Linked Articles
eeNews Europe