Reduce energy consumption in data centers

The capital expenditure and maintenance are the major expenses in the ICT infrastructure, and the energy consumption is emerging to be the main opex cost. It is estimated that energy constitutes around 10% of an IT budget currently, and could account for 50% of the average IT budget in just a few years.

While the world has less energy to spare, the need for greater energy-efficient performance is becoming essential. An area, where the overall difference can be made is at the desktop level and client devices, consuming 39% of the overall power requirements of the complete IT deployment.

The Core Solution

As the trend moves within the data center, from single-core to multi-core processors in the computing performance leap, multi-core architecture has a significant role in reducing energy consumption. The multi-core architecture enabled run demanding applications, while simultaneously maintaining system responsiveness-

• Keeping systems protected, efficient, and manageable while minimizing impact on productivity
• Providing headroom for future business growth and new solution capabilities

In a data intensive application environment such as SCADA, multi-core technology can boost the performance, by enabling a software developer to divide the application into multi-threaded data streams. Since each execution core operates as a separate processor, the multi-core processor provides system designers with the option of running up to multiple applications on a single chip.

As the industry moves to multi-core processor environment, system architectures have several options for the organization, one of the most important system resources-the cache. The multi-core architecture allows both cores to have access to the entire cache, reducing possible resource under-utilization. The enhanced data pre-fetch logic improves efficiency by pre-fetching data to the cache even before cache requests occur.

However, some architectures choose to keep the last-level cache private to each core for simplicity, while other architectures explore sharing the last-level cache among different cores for better performance/cost ratio and improved resource allocation. The shared cache is dynamically allocated to each processor core, based on workload. This significantly reduces latency, improving performance. This shared cache in the latest processor architecture also features a new power-saving mechanism that enables the cache to dynamically smooth its ways into system memory, based on demand, or during periods of inactivity.

 

The benefits of Shared Cache System:

• Efficient usage of the last-level cache is useful because if one core idles, the other core takes all the shared cache. This reduces resource under utilization
• Flexibility for programmers allows more data-sharing opportunities for threads running on separate cores that are sharing cache where one core can pre/post-process data for the other core while enabling alternative communication mechanisms between cores
• Reduced cache-coherency complexity leads to reduced false sharing because of the shared cache and the workload is lesser to maintain coherency when compared to the private cache architecture
• Reducing data-storage redundancy prevents duplication of data
• Reduce front-side bus traffic allows effective data sharing between cores which lets data requests to be resolved at the shared-cache level instead of going all the way to the system memory

Conclusion

The multi-core architecture needs to be supplemented with a number of innovations in micro architecture to achieve significant performance-per-watt improvements. Transistor size and structure are at the very center of delivering the benefits of Moore's Law to the end user-the smaller and more power efficient the transistor, the better.

Therefore multi-core technology supplemented with advances in micro architecture is enabling an innovative form of factor designs in PCs, tablets, and phones with ever increasing power efficiency. As the data and technology penetration expands in the world, energy efficiency achieved through these technologies is extremely important to reduce power consumption in not only ICT industry and data centers, but also in all other industries such as infrastructure, energy, logistics, and transportation to name a few.