This is the third article in a series taken from the insideHPC Guide to Flexible HPC.
Organizations that implement high-performance computing (HPC) technologies have a wide range of requirements. From small manufacturing suppliers to national research institutions, using significant computing technologies is critical to creating innovative products and leading-edge research. No two HPC installations are the same.
For maximum return, budget, software requirements, performance and customization all must be considered before installing and operating a successful environment.
Requirements in HPC Environments
Organizations that require HPC technologies will have a wide range of requirements, depending on their workloads. While some applications can run hundreds or even thousands of cores, other applications cannot take advantage of more than one core.
A number of industries require HPC technologies, to either bring products to market faster or model and simulate physical phenomena.
BIO-INFORMATICS, LIFE SCIENCES and COMPUTATIONAL CHEMISTRY
For these industries, the advancement in computational power has enabled companies to offer new drugs and medicines by simulating how certain components interact with other components. The computational power available in a single server containing multiple CPUs and accelerators brings personalized medicine closer than ever before. This is definitely an industry that can take advantage of the wide range of hardware that is available today.
ELECTRONIC DESIGN AUTOMATION
Designing the latest computer processors and other microelectronic devices requires significant computing power. While not an overly parallel application, millions of simulations must be run in order to verify the inner workings of a new CPU or related device. Large clusters of low core-count systems would be ideal for this industry so that runs can all be done in parallel, thus reducing the chance for expensive hardware bugs. Bringing a new product to market faster can increase profits and result in a market-leading product family.
The ability to make faster decisions with more data can directly increase the profits of a trading organization. With the latest systems, more simulations based on previous data and trends can be run, increasing the confidence in the outcomes. Trading decisions can be made milliseconds faster than a competitor whose systems (clock speed) may be older, leading to a competitive advantage.
Large-scale computer simulations have been driving the competitive edge for manufacturing companies for many years. For accurate simulations that involve both structural and computational fluid dynamics, they are ideally run in large clusters with varying numbers of cores. Recent advances in commercial computational fluid dynamics (CFD) codes can take advantage of hundreds of cores to produce more accurate and more complete simulations. While some critical applications that a large manufacturing company will run are not highly parallel, a large cluster with varying types of systems allows for multiple design teams to share computer resources. This results in a more optimized product and faster time to market.
DATA MINING, ANALYTICS and DATABASES
Although not traditional HPC, the use of significant computing power for industries that must sift through terabytes of data is required. High-performing CPUs are needed, as is a lot of high-speed main memory to hold the massive amounts of data. Some of these tasks can be run on a cluster using a variety of computing hardware.
WEATHER, CLIMATE MODELING and ATMOSPHERIC RESEARCH
HPC systems are used every day to simulate near-term weather events or long-range climate forecasts. Finer meshes with more physics simulated can lead to more accurate day-to-day or even hour-to-hour weather forecasts. Long-range climate modeling also benefits from a large amount of computing power. In addition, weather and climate simulation codes can take advantage of thousands of cores and are highly scalable. Many of these applications can also use the power of accelerators to increase performance significantly.
OIL, GAS and PETROLEUM EXPLORATION
Making correct and accurate decisions when deciding where to drill for oil and gas can save millions of dollars. HPC technologies can aid in determining where to drill and can simulate the best methods for oil and gas extraction. Simulating these processes can take advantage of all the computing power that an organization can budget for. Large clusters of a wide range of hardware can be effectively used to look for valuable deposits, which reduces the expense at the front end of the exploration pipeline.
Many disciplines that are being researched at universities, national labs and in the research departments of leading companies rely on highly optimized applications and a wide range of computer systems. Together, large clusters can be created that allow researchers to simulate many processes or search for new answers. Many of these applications are highly parallel and are custom-designed. These applications can take advantage of the latest in CPU architectures and heterogeneous environments. Tweaking the system to perform at its maximum performance for a specific application requires extreme knowledge of the application, as well as a deep understanding.
In coming weeks, this series will consist of articles that explore:
- Challenges in HPC Environments
- Requirements in HPC Environments
- TYAN® Solutions
- Successful Implementations