From Apple Learn & Earn course "Understanding Compute-Intensive Problems & Solutions&am
A high-performance computing solution is needed when the amount of time required for a single task to complete becomes unacceptable.
Many scientists are trying to understand how a linear molecule folds itself into a three-dimensional structure. Studying how a protein folds provides insight into how proteins work--such as how hemoglobin is able to fold and surround an iron molecule in order to carry oxygen through the body. Protein folding experiments are very computationally intensive. A single protein folding experiment on a 1 GHz Power Mac G4 can easily take several days. It would be ideal to speed up this process to a more reasonable amount of time, say 1 hour. This would require an HPC solution. Currently, the most cost-effective method of creating an HPC solution is with a cluster of off-the-shelf and readily available computers. This type of cluster accelerates the process by working cooperatively on the same computationally intensive task. The task of computational protein folding is considered tightly coupled, since each part of the simulation affects other parts of the simulation.
__________________________________
# Meta computing. This solution involves the use of available distributed computing resources for high-throughput solutions. Familiar examples are SETI and Folding@Home--Stanford University's (Pande Group) distributed simulation project to determine how proteins assemble themselves. These solutions require specialized application software, and the volume of processing depends entirely on the number of resources available.