Zucchini : ECHELON Scales To Multiple GPUs To Further Accelerate Simulations

Learn more about ECHELON's scalability test for this giant field model in the Middle East.

Bg case study zucchini

Jan 25 2022

A giant Middle East asset seamlessly runs on a single GPU using ECHELON software and is further accelerated by using 4 GPUs.

By the numbers
0 M
Active Cells
0 +
Wells
0
Years of Simulation
Stone Ridge Technology – ECHELON Advantages

Challenge

Zucchini is a giant Middle-Eastern field with a long production history, yet with a low recovery factor and on the order of 40 more years of production into the future. While treated as black-oil, the model has some complexities including both relative permeability and capillary pressure endpoint scaling and analytical aquifers. The grid block size is 100x100x2 meters, for a total of 6.7M cells. The challenge is to run the forecast with a sequential drilling queue for the management of infill wells, in a robust-enough way to allow optimization of the development plan without being impacted by field scheduling instabilities.

Results – Stone Ridge Technology
Cross-section of the Zucchini field model, colored by porosity

Solution

Using ECHELON one can seamlessly run giant reservoir models on a single GPU with exceptional performance; in addition, ECHELON can be further accelerated by scaling to multiple GPUs using domain partitioning as illustrated in the Figure (1).

Solution – Stone Ridge Technology
Figure 1. Partition of the Zucchini field in 4 domains, generated by ECHELON for the 4-GPU run

Results

The number of operating producers (Figure 2, top) is almost unaffected by the number of GPUs, including when the drilling queue is active towards the end of the production plateau, confirming the robustness of ECHELON. The scalability study (Figure 2, bottom) shows that using 4 GPUs yields a 2.6x speed-up compared to a single GPU.

Results – Stone Ridge Technology
Figure 2. (Top) Number of active producers using four different number of GPUs for the simulations, and field oil production rate. The simulation spans both history and forecast periods. (Bottom) Scalability test for the Zucchini field model, on NVIDIA V100 GPUs. Scalability is close to ideal from 1 to 2 GPUS (At 1.8x), and progressively decreases as more GPUs are added. Using 4 GPUs provides a 2.6x speedup compared to the single GPU case.

We thank Eni S.p.A for the permission to publish the data contained in this case study.

Download Case Study

Additional case studies

Deep dives into our software product and its efficacy.

ECHELON: A Cloud Computing Use Case – Stone Ridge Technology

ECHELON: A Cloud Computing Use Case

Evaluating the productivity enhancements of Fishbones in fractured carbonate reservoirs using high-performance cloud-based reservoir simulation.

View case study →
Unconventional Field Development – Stone Ridge Technology

Unconventional Field Development

Workflow to manage the operational challenges in an unconventional field development requires full field simulation.

View case study →
Kabosu: ECHELON Software Tackles Complex Field with Flying Colors – Stone Ridge Technology

Kabosu: ECHELON Software Tackles Complex Field with Flying Colors

ECHELON software tackles an extremely difficult compositional model. A demonstration of the robustness of ECHELON software.

View case study →