QMCkl Code
CNRS
University of Twente
SISSA School
University of Versailles Saint Quentin en Yvelines

QMCkl

The QMCkl library aims at providing a high-performance implementation of the main kernels of Quantum Monte Carlo methods.

Anthony Scemama (CNRS), Claudia Filippi (UT), William Jalby (UVSQ), Saverio Moroni (SISSA), Pablo Oliveira Castro (UVSQ), Sandro Sorella (SISSA), Cedric Valensi (UVSQ)
As part of TREX: Vijay Gopal Chilkuri (CNRS), Francois Coppens (UVSQ), Evgeny Posenitskiy (CNRS)

The QMCkl library aims at providing a high-performance implementation of the main kernels of Quantum Monte Carlo methods. A first implementation of the library focuses on the definition of the API and the tests, and on a pedagogical presentation of the algorithms. Then, HPC experts can use this initial implementation as a reference to re-write optimized versions of the library with the same API.

Audience

Theoretical chemistry community, quantum simulation of materials community.

Video Interview with Anthony Scemama from CNRS

 

NEED HELP WITH THIS CODE? GET SUPPORT

Current Stage and Next Step

The first version of the pedagogical library was released, with the possibility to compute the local energy for single-determinant wave functions using a basis of Gaussian functions and the Jastrow factor implemented in CHAMP. The next step in the pedagogical library is the computation of multi-determinant wave functions, and the implementation of arbitrary atomic basis sets. In parallel, we will start the development of a high-performance implementation of the library, adapted to GPU accelerators.

 

NEED HELP WITH THIS CODE? GET SUPPORT