Quantum ESPRESSO toward the exascale

Area of Science:

• Computational Materials Science
• Quantum Mechanics
• High-Performance Computing

Background:

• Quantum ESPRESSO is a widely used open-source software for quantum-mechanical materials modeling.
• It relies on density-functional theory, pseudopotentials, and plane waves.
• Current hardware architectures face energy constraints for exascale computing.

Purpose of the Study:

• To present the motivation for porting Quantum ESPRESSO to heterogeneous architectures.
• To review the ongoing efforts in adapting the code for hardware accelerators.
• To address energy efficiency challenges in materials modeling for exascale.

Main Methods:

• Porting the Quantum ESPRESSO code to heterogeneous computing architectures.
• Leveraging hardware accelerators (e.g., GPUs, FPGAs) for computation.
• Benchmarking performance and energy consumption on diverse hardware.

Main Results:

• Demonstration of Quantum ESPRESSO's adaptability to hardware accelerators.
• Identification of performance gains and energy savings through acceleration.
• Validation of the approach for future exascale systems.

Conclusions:

• Porting Quantum ESPRESSO to hardware accelerators is crucial for advancing materials modeling.
• This strategy will enable more energy-efficient and powerful simulations.
• The work paves the way for achieving exascale computing in materials science.