A Comparative Analysis of OpenMP and Cuda in Real-Time Applications
Devesh Nahar, Mangalya D. Phaye, Ridima Garg, Anju Susan George, Chithra K · 2025
The need for effective parallel computing in real-time applications has gained momentum with advancements in multi-core processors and GPUs. OpenMP and CUDA are prominent frameworks, with OpenMP streamlining multi-core CPUs and CUDA utilizing GPU parallelism. This comparison measures performance in real-time applications like image processing, scientific computing, and medical diagnostics. The results show that CUDA performs 2.5×–16× speedups for highly parallel applications, and OpenMP is competitive in memory-bound workloads. This can inspire industrialists and research scholars to implement CUDA for sustainable development and economic growth in parallel processing systems. Chief challenges, optimizations, and hybrid implementation possibilities are elucidated.