Accelerating lattice QCD simulations with 2 flavors of staggered fermions on multiple GPUs using OpenACC—A first attempt

Abstract

We present the results of an effort to accelerate a Rational Hybrid Monte Carlo (RHMC) program for lattice quantum chromodynamics (QCD) simulation for 2 flavors of staggered fermions on multiple Kepler K20X GPUs distributed on different nodes of a Cray XC30. We do not use CUDA but adopt a higher level directive based programming approach using the OpenACC platform. The lattice QCD algorithm is known to be bandwidth bound; our timing results illustrate this clearly, and we discuss how this limits the parallelization gains. We achieve more than a factor three speed-up compared to the CPU only MPI program.

Acceleration of three-dimensional Tokamak magnetohydrodynamical code with graphics processing unit and OpenACC heterogeneous parallel programming

“…Compared with the CUDA and OpenCL which require great efforts on code redevelopment, OpenACC has many advantages, such as satisfactory acceleration with very few modifications on an original source code and good compatibility with other devices, for example central processing unit (CPU). It has been successfully applied in some scientific and engineering codes, such as the flow code NeK5000 (Markidis et al 2015), the computational electromagnetics code Nekton (Otten et al 2016), the C++ flow solver ZFS (Kraus et al 2014), the Rational Hybrid Monte Carlo (RHMC) QCD code (Gupta and Majumdar 2018), the Gyrokinetic Toroidal Code (GTC) (Wang et al 2016), the space plasma Particle-in-cell 3 / 31 (PIC) code iPIC3D (Peng et al 2015), the three dimensional pseudo-spectral compressible magnetohydrodynamic GPU code G-MHD3D (Mukherjee et al 2018), the solar MHD code MAS (Caplan et al 2018), and etc. The application perspective of the OpenACC technology in other scientific and engineering areas is very good and attractive.…”