High-performance computing

High-performance computing uses powerful computers and coordinated processing to solve problems that demand more computation, memory, or data movement than ordinary systems can provide.

Parallel computing: Parallel computing divides work among multiple processing units that execute tasks simultaneously. HPC gains speed by distributing suitable workloads across many processors.
Numerical weather prediction: Numerical weather prediction uses mathematical models and observations to calculate future atmospheric conditions. Forecast models require extensive computation to evolve atmospheric states across space and time.
Computer architecture: Computer architecture describes the structure and operation of a computer system, including its processors, memory, and communication. HPC performance depends on how these components are organized and connected.
Cloud computing: Cloud computing delivers configurable computing resources over networks, typically with usage-based provisioning. Cloud platforms offer flexible capacity, while traditional HPC centers emphasize tightly coupled systems and specialized workloads.
TOP500: TOP500 is a project that ranks the world's most powerful supercomputer systems using the LINPACK benchmark. Its rankings make headline computational performance comparable across major systems.
Message Passing Interface: Message Passing Interface is a standard library specification for communication among processes in parallel programs. It lets distributed HPC processes exchange data across nodes.
Molecular dynamics: Molecular dynamics simulates the movement and interactions of atoms and molecules over time. Large simulations use HPC to model many particles and long timescales.
Cache coherence: Cache coherence ensures that multiple processor caches maintain a consistent view of shared memory. Shared-memory HPC programs need consistent data when processors cache the same locations.
High-throughput computing: High-throughput computing maximizes the number of independent jobs completed over time. It prioritizes aggregate job completion rather than the fast execution of tightly coupled calculations.
Green computing: Green computing seeks to reduce the environmental impacts of computing through design, operation, and disposal choices. Large HPC facilities consume substantial electricity, making efficiency a central concern.

Linked from 21 pages