How does HPC work?
Scientists, engineers, and researchers use high performance computing (HPC) for incredibly computationally intensive tasks, like those used in weather forecasting, oil and gas exploration, physics, quantum mechanics, and other areas.
While HPC can be run on a single system, its real power comes from connecting multiple HPC nodes into a cluster, which may also be called a supercomputer. This provides the capabilities to compute extreme-scale simulations that are not feasible or even possible on a single system.
Some of the first and most prominent supercomputers were developed by Cray, which today is an Intel® Datacenter Builders partner. Modern supercomputers are large-scale clusters made up of CPUs, accelerators, high-performance communication fabric, and sophisticated memory and storage, all working together across nodes to prevent bottlenecks and deliver the best performance.
Scaling up performance
HPC applications take advantage of hardware and software architectures that spread computation across resources within a single system, typically a server. This form of parallel processing is called “scale-up.” Performance gains are powerful but are limited to the capabilities within a single system.
Scaling out performance
Additionally, HPC applications can take advantage of spreading computation across multiple systems that are configured to act as one system. This is called an HPC cluster, and it enables applications to spread computation running in parallel across a number of systems.
Cost and benefits of HPC
HPC is growing in the enterprise space. It’s also becoming more accessible. Scientists and engineers can run HPC workloads on their on-premises infrastructure, but they can also scale computing power without large capital expenses by using resources from public cloud service providers.
HPC use cases
Today, businesses are using HPC for simulation and modeling in key areas, such as developing autonomous driving models, designing and manufacturing products, analyzing seismic waves and their effects on structures, and evaluating oil and gas production strategies. HPC systems are also advancing precision medicine, financial risk assessment, fraud detection, computational fluid dynamics, and other areas.
The most productive HPC systems feature a highly tuned combination of hardware and software. Hardware for HPC usually includes high-performance CPUs, fabric, memory and storage, and networking components, plus accelerators for specialized workloads. HPC platform software, libraries, optimized frameworks for big data and deep learning, and other software tools help you get the most from your HPC clusters.
HPC and Intel
Intel delivers a comprehensive portfolio of hardware and software to meet the needs of high performance computing. This is HPC on Intel.
It all starts with Intel® Xeon® Scalable processors, which provide a foundation for breakthrough HPC and a leap in agility and scalability. It continues with networking options like Intel® Ethernet products and Intel® Omni-Path Fabric (Intel® OP Fabric) for scaling optimizations. Intel® Optane™ DC technologies offer innovations for memory and storage, and Intel® FPGAs provide acceleration options to speed certain workloads. Intel® software and developer tools provide the best path to performance to take advantage of Intel® architecture and technologies.
With this broad range of products and technologies, Intel tackles HPC both big and small. Intel is helping achieve exascale computing for the world’s most demanding scientific challenges and provides a standards-based approach for solutions aimed at common HPC application workloads through the Intel® HPC Platform Specification. Accessibility is growing, and Intel is helping users realize the benefits of advances in the technologies that support HPC.
The convergence of AI and HPC
The combined power of HPC and machine and deep learning shows amazing promise across a variety of applications, from linguistics to genomics sequencing to global climate modeling. Intel is powering the convergence of AI and HPC by working with ecosystem partners to develop reference architectures and solutions specially designed to blend these practice areas.
Intel® Select Solutions for HPC
Finding the right combination of hardware, software, and configurations for HPC can be a challenge. Explore Intel® Select Solutions for HPC for a fast, easy way to match application needs with solutions for simulation and visualization, simulation and modeling, and genomics analysis.
Driving advances in HPC
The latest advances in computing and software are resulting in leaps forward for HPC. At Intel, we’ll continue to work closely with the HPC ecosystem to optimize hardware performance; blend HPC, AI, and analytics; and forge the best path to exascale computing.
With high performance computing, you get power in numbers, as many connected HPC nodes can perform calculations in parallel.