Parallel Computing
A computational architecture where large physical calculations are divided into smaller sub-tasks executed simultaneously across multiple processor cores or GPUs.
Governing formula
Amdahl's Law Speedup S(P) = 1 / ((1 - p) + p / P)
SI unit
Speedup Factor S(P)
In depth
Utilizes Message Passing Interface (MPI) and CUDA GPU acceleration to solve massive 3D grid simulations. Amdahl's law specifies maximum speedup bounds based on non-parallelizable sequential fraction (1-p).
Examples in the real world
Running 3D cosmological hydrodynamics simulations on supercomputers using thousands of parallel CPU/GPU cores.