Case Study: Zucchini achieves faster giant reservoir simulations with Stone Ridge Technology’s ECHELON

Zucchini Echelon Scales to Multiple Gpus to Further Accelerate Simulations

Stone Ridge Technology worked with Zucchini, a giant Middle Eastern oil field with a long production history, low recovery factors, and around 40 more years of production remaining. The team needed a robust way to run large black-oil reservoir forecasts with a sequential drilling queue for infill well management, while avoiding scheduling instabilities in a model with 6.7 million cells, more than 200 wells, and complex features such as relative permeability and capillary pressure endpoint scaling and analytical aquifers.

Stone Ridge Technology implemented ECHELON 2.0, its GPU-based reservoir simulator, to run the field model efficiently on one or more NVIDIA V100 GPUs. The solution handled the giant simulation on a single GPU and scaled across multiple GPUs using domain partitioning; performance was close to ideal from 1 to 2 GPUs at 1.8x, and 4 GPUs delivered a 2.6x speedup versus a single GPU. This enabled Zucchini to run century-long simulations more quickly and reliably for development planning and optimization.