3 min read

Will 2020 be the year your company adopts High Performance Computing?

10 February 2020

HPC is revolutionising what companies can do with their data analysis. It's being adopted at a fast pace, and to help you figure out if HPC - as it stands right now - is a good fit for your company, we've put together a brief history and a review of HPC's current status. 

[Researching your next IT upgrade? Learn all about colocation here]


A recent study of UK IT professionals found:

  • 74% believed that HPC could positively impact their business
  • 43% said their business was already using a practical application of HPC
  • 48% were considering it

These figures are high, but that’s no great surprise. We’re living in the age of Big Data and all the data in the world is not much use unless you can process it in a reasonable amount of time. HPC enables businesses to do just that – and so much more. Whether it’s data processing, scanning for patterns or insights, or running simulations that help you create ground-breaking new products, aggregating CPUs and GPUs into an HPC cluster can bring you big benefits.

Before we look at some examples of how businesses are taking advantage of HPC, let’s take a brief trip down memory lane to see where it all began.

[If you're considering HPC for your business contact 4D to explore our HPC solutions]

1964 – The dawn of the supercomputer

We can’t talk about HPC without mentioning supercomputers. Supercomputers could be described as the mother of High Performance Computing: the pinnacle of computing capabilities. Not necessarily practical for your average SME, but interesting nonetheless.

The first supercomputer was the CDC 6600, which held the title of the fastest computer in the world from 1964 to 1969. It was used in research laboratories for endeavours like analysing nuclear events.

From the 1960's, supercomputers continued to increase in both capacity and capability, but their role was largely in the sphere of scientific research, for example: weather forecasts, mapping out your best bets for oil and gas exploration, and trying to understand the origins of the universe.

These computers cost millions of dollars and in the early days they were dealing with the kind of data that was only available/interesting to the scientific experts who spent long days in dimly-lit laboratories. (I’m generalising, of course. The lighting was probably very good.)

But these days there is a wealth of data available to just about everyone – ‘wealth’ being the operative word. Enterprises around the world are mining that data for valuable insights, which they put to use to improve their services and develop new tools and technologies.

How are enterprises using HPC today?

Data processing and analysis is one of the main drivers of the enterprise uptake of HPC. The use of artificial intelligence (AI) and machine learning (ML) to organise and gain insights from all this data requires the greater processing power and speed of HPC. Business sectors as diverse as financial services and medical research are taking advantage of the capacity and capabilities of HPC to do things like detect fraudulent activity on an account or predict flu outbreaks.

Meanwhile, on the engineering side, the capacity for advanced modelling and simulation enables breakthroughs in product development, as well as significant performance gains in existing products and processes. (If you’re currently trying to run these kinds of workloads on a desktop with GPU you can imagine how the greater processing power of HPC might help speed things up!) And then you have film and television studios making use of giant render farms – i.e. HPC – to render computer generated imagery.

Let’s look at a couple of specific use cases:

  • TGen – This US-based non-profit genomics research institute uses HPC to help identify genomic discoveries that improve diagnostics, prognostics and treatments for a range of disorders. The institute was using millions of CPU hours for data analysis. Results could take weeks – time that desperate patients couldn’t afford to lose. With the adoption of HPC, workloads were reduced from weeks to hours, buying more time for doctors to help patients.

  • Paddy Power – The bookmakers turned to HPC to enable them to turn data into odds. With so much data coming in that could affect the outcome of each event, Paddy Power needed the massively parallel processing capacity of HPC to make complex, real-time calculations as quickly as possible.

  • MiNO Marine – HPC enabled MiNO Marine to design a new type of tow boat in less than 12 months. With the ability to simulate the movement of both fresh and seawater, water pressure, wave impact and other factors, the team was able to design a safe, efficient, high-performance vessel in a relatively brief amount of time. These simulations would have been, at best, slow on desktop machines, but with HPC multiple simulations could be run at one time.

You’ll notice the common denominator here: time. Whatever you are trying to achieve can be managed much more quickly with HPC.

The physical factor

Adding processing capacity and capability means adding to the physical dimensions of your computer. More processors; more cores; a significantly greater power requirement; more storage space needed and an increased risk of something going wrong.

Unless you have the appropriate facilities in house, it is recommended you store your HPC in an HPC-ready data centre where storage space is not an issue and where the environment is perfectly adapted to the power and cooling needs of high performance computers.

Traditional medium density CPU storage will require a power deployment around 3 to 7 kW per rack.

HPC clusters of CPU and GPU cores demand 20 to 40 kW per rack, increasing the heat load drastically.

At 4D Data Centres, we have invested in rear-door cooling to bring those temperatures back down to typical levels. Our system uses a cooling loop on the back of the rack, which is connected to the same chilled water system that feeds the main data floor CRAC (Computer Room Air Conditioning) units. This means we can still offer high density storage for HPC clusters, safely and efficiently.

What is the future for HPC in business?

If we revisit this article even 12 months from now, I expect we will be able to find many more use cases for enterprise adoption of HPC. Businesses of all sizes will want to take advantage of the advances made possible by AI and ML – and they won’t be able to do that without upgrading their hardware capacity.

If you are making plans to adopt HPC in your company, make sure you also consider how you intend to protect the investments you make. Power deployment, cooling, safety and security are all critically important to the productivity and performance of your business. Allocating those tasks to a HPC-ready data centre gives you the freedom and the peace of mind to crack on with your next big breakthrough.

4D's HPC Capabilities

Across the UK, our data centres have been uniquely built for high density deployments, and our infrastructure reflects this.   All our racks are equipped with power supplies capable of hosting anything up to ultra-high density (around 40kW) without any additional power installation requirement.  The latest rear-door cooling technology comfortably copes with the cooling requirements of a HPC rack and can be installed upon request onto any of our existing racks.

4D’s high and ultra-high density capabilities provide a leading-edge HPC standard that can seamlessly integrate across other systems and services.  If you’re considering hosting HPC within a data centre, please get in touch with 4D or call us on 020 3962 0399 to explore our HPC solutions and how we can best help your company.

New call-to-action