Calculations on Graphic cart

In collaboration with an industrial drilling partner and a leading university, we migrated a CPU-based thermodynamics simulation application to GPU computation using CUDA. To support this transition, we assembled two dedicated workstations:

• 1. Machine 1 featuring an GTX 4080 GPU
• 2. Machine 2 equipped with five GTX 3070 GPUs

By fully porting the simulation algorithm from CPU to GPU, we achieved a 40-fold reduction in computation time—what previously took two weeks to compute now completes in just six hours. This breakthrough significantly accelerates the development and validation.

Significant Speedup: Reduced computation time from two weeks to six hours by leveraging CUDA parallel processing on GPUs.

Enhanced Workflows: Faster simulations allow more frequent iteration and experimentation, expediting the development of drilling solutions.

Collaborative Success: Joint effort with industry and academia ensured real-world relevance and scientific rigor in the simulation models.