Last year Clabby Analytics described what we saw as a major shift in the high performance computing (HPC) market. For years HPC was largely associated with large-scale scientific workloads (for example: pharmaceutical research) and with applied technical computing where it was used primarily for engineering and simulations in electronic design automation, automotive/aerospace engineering and petroleum discovery/analysis. In late 2010 and throughout 2011, however, we noticed a shift in the HPC market as new workloads such as digital media, various financial services applications, new life sciences applications, on-demand cloud computing services and analytics workloads made their way onto HPC servers.
We are now seeing another new trend developing in the HPC space with the introduction of ultra-dense servers. The arrival of ultra-dense, hyper-scale designs such as IBM’s NeXtScale is changing the system design dynamics in the HPC marketplace. NeXtScale is based on standard components (such as standard racks, memory, and other components). These servers can be easily serviced (from the front). They are pretested for reliability. They are optimized for performance and efficiency. In this whitepaper from IBM, we learn how these types of ultra-dense systems have been designed to serve scale-out applications that require high computing throughput such as technical computing, cloud, grid and analytics workloads.
Server Density is important from for two reasons:
- It reduces the amount of real estate that a system occupies (footprint); and,
- It reduces the amount of supporting infrastructure (networking components, management facilities, etc.) needed to support the system design.
Learn how a NeXtScale platform can help your organization:
- Deploy twice as many servers per floor tile;
- Use less energy;
- Reduce cooling costs (because air and water cooling can be used – and water is thousands of times more efficient as a conductor than air);
- Improve time-to-result (the speed at which solutions are derived) due to processing efficiency and low communications overhead;
- Improve system reliability; and,
- Reduce overall systems cost by purchasing industry standard components (such as industry standard racks and memory).
Remember – Enterprises looking for raw computing power and raw throughput to handle technical computing, cloud, grid and analytics would be well served to evaluate this new type of ultra-dense, highly scalable systems design.
Download this whitepaper today from the insideHPC White Paper Library.