Full text loading...
-
Solving the Reservoir Simulation Bottle Neck with a Hybrid Approach Based on Modern CPU and GPU Computational Platforms
- Publisher: European Association of Geoscientists & Engineers
- Source: Conference Proceedings, 80th EAGE Conference and Exhibition 2018, Jun 2018, Volume 2018, p.1 - 5
Abstract
The main objective of the modern reservoir simulators is to provide a realistic description of the reservoirs, fluids and hydrocarbon extraction technology in order to achieve a reliable and accurate predictions, guarantee excellent performance and parallel scalability.
In the past, the simulation performance was largely dependent on the memory throughput of CPU based computer systems. Such a limitation has recently been improved when the new generation of graphical processing units (GPU) became available for general purpose computing.
In this abstract we discuss challenges and developed solutions for running reservoir simulations using modern CPU+GPU hardware architecture and propose a methodology to distribute the workload between various parts efficiently (hybrid approach). The approach is tested on several data sets and run on the full physics reservoir simulator on various computational platforms, such as personal computers and clusters with and without GPU’s involved.
The proposed technology demonstrates multifold speed up for models with substantial number of active grid blocks. In some cases, the speedup obtained by the hybrid approach can be as high as 3–4 times compared to the traditional GPU-based approach. Considering the recent progress in the GPU development, this factor is still expected to grow.