<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=1875728222659097&amp;ev=PageView&amp;noscript=1">

HPC's Time Of Change: The Many-Core Era & Accelerated Computing

Home » Blog » HPC's Time Of Change: The Many-Core Era & Accelerated Computing

A TIME OF CHANGE FOR HPC: INCREASING PERFORMANCE IN 'THE MANYCORE ERA'

Change is nothing new in the world of high performance computing (HPC). In fact, over recent decades, the industry has witnessed a number of developments. From the introduction of x86 clusters in the late 1990s to the decade-dominance of vector processing, industry trends and adaptations are simply reflective of technological advances and consumer demand.

Currently, the world of HPC is experiencing a transition into a new era – the many-core era. This era, like those that came before, is driven by cost, performance and accessibility, with the numerous cores producing more computational power.

Over recent years, the use of accelerators has become a significant part of this transitional period. According to the latest research, approximately a third of HPC systems in operation today are using accelerators, and for the first time, over 100 accelerated systems are featured on the Top 500 list. In addition, a third of FLOPS are powered by accelerators.

A key factor in the increase of accelerator-use is the continued slow-down of Moore’s Law. With the size of transistors approaching atomic scale, improving microchip performance without experiencing a disproportionate increase in costs and power is difficult.

2016-07-26 - McLeod Engg on Tap.png

This slow-down has meant that the industry can no longer rely on witnessing a doubling of performance every 18 months, and for this reason the adoption of accelerators has become commonplace. The industry has recognized the increased efficiency that even a smaller accelerator investment can make, so their adoption is not surprising.

The most widely used accelerator type is easily the GPU, holding holding around 80% of the market. As the accelerator type that brought supercomputing into everyday life, the mobile phone is a good example of its working existence.

Envenio predicts that accelerator use will continue to increase, and within a few years, a larger majority of Computational Fluid Dynamics (CFD) systems will already be accelerator equipped. 

PROACTIVE OR REACTIVE?

Alarmingly, two-thirds of companies are currently predicting that they will require a 10x increase in modelling performance over the next five years. To keep up with performance increase expectations, CFD software users face two choices - to be reactive or proactive.

By being reactive, CFD simulation teams continue to grow their computing cluster, eventually running into limitations and restrictions such as power, space, additional users and license costs. Over recent years, a number of advances have been seen in the development of hardware, but CFD vendors have been less proactive when it has come to accelerator-supported HPC software packages. Their platforms only achieve marginal gains from using HPC technology, which means the benefit does not justify the additional HPC hardware cost. 

By being proactive, the industry embraces and invests in true HPC enabled software, ultimately limiting and even reducing the amount of hardware required, and providing a platform from which they can grow and evolve with future trends. While writing code for accelerators is not easy, this is what was required to arrive at a truly optimized solver for HPC hardware. Envenio built its EXN/Aero software package to deliver economical supercomputing for all sizes of CFD teams (freelancers to large corporate teams), allowing businesses to not just reduce hardware investment, but to be confident in their ability to increase performance, and ultimately remain in line or ahead of their competitors. 

Envenio is already looking forward to the many-core era, identifying barriers and finding ways to overcome and adapt to them. A current example of this is EXN/Aero’s development of space-time parallelization, which will help to overcome the new latency barriers experienced in a many-core environment. Built to not only serve the needs of today, the platform is designed in such a way that it is ready to accommodate future trends that arise from the exa-scale installations and developments.

The real question for CFD users is whether you will be reactive, or proactive?


  

2016-11-4 | Categories: CFD, HPC

Popular Posts