Improving Performance and Energy Efficiency of Geophysics Applications on GPU Architectures

Abstract

Energy and performance of parallel systems are an increasing concern for new large-scale systems. Research has been developed in response to this challenge aiming the manufacture of more energy efficient systems. In this context, this paper proposes optimization methods to accelerate performance and increase energy efficiency of geophysics applications used in conjunction to algorithm and GPU memory characteristics. The optimizations we developed applied to Graphics Processing Units (GPU) algorithms for stencil applications achieve a performance improvement of up to 44.65% compared with the read-only version. The computational results have shown that the combination of use read-only memory, the Z-axis internalization and reuse of specific architecture registers allow increase the energy efficiency of up to 54.11% when shared memory was used and increase of up to 44.53% when read-only was used.

Energy efficiency and portability of oil and gas simulations on multicore and graphics processing unit architectures

“…They reached performance improvements of up to 10×. Pavan et al 37 proposed architecture‐based optimizations to improve the energy efficiency of geophysics applications running on NVIDIA GPUs. Their results showed that a combination of shared memory and reuse of registers improve performance and energy efficiency by up to 54.1%.…”

“…Pavan et al 37 and Calore et al 39 proposed different GPU optimizations to a stencil code evaluating performance and energy efficiency. We go beyond by assessing the portability of a real‐world application for different architectures and not only GPUs.…”