HPC Acceleration

In the ever-evolving landscape of engineering simulations, speed and accuracy are paramount. Barracuda Virtual Reactor has been meticulously crafted to harness the power of parallel computing, delivering exceptional performance gains over traditional CPU serial simulations.

CPFD Software’s engineers spotted the potential for GPU acceleration over a decade ago, and with our first GPU release in 2014, we immediately saw a 5-10x increase over our CPU speeds. Since then, with hardware advancements from NVIDIA as well as moving more calculations to parallel, this number has spiked tremendously. Our software has continued to see extensive parallelization with the introduction of multi-GPU parallelization, resulting in simulation speeds that are 50x to 500x faster than CPU serial speeds.

More Thorough R&D via Broader Parameter Space Searching: The increased simulation speeds empower researchers to explore a vastly expanded parameter space. This means more comprehensive and nuanced investigations, leading to detailed studies and more robust conclusions.

Higher Complexity Models with Increased Fidelity: With Barracuda Virtual Reactor’s enhanced speeds, engineers and scientists can now incorporate more complex models without sacrificing fidelity. This opens the door to simulating intricate scenarios with a level of detail previously deemed impractical.

Solving Previously Intractable Problems: The software’s unprecedented speeds enable the resolution of problems that were once considered computationally intractable. Tackle challenges that were previously beyond the reach of simulation, pushing the boundaries of what’s possible in fluid dynamics research.