Ramboll cuts simulation processing time by 85 percent

Driving business growth with high-performance computing from IBM and Gridcore

Published on 18-Apr-2013

"Thanks to IBM and Gridcore, we can process more detailed simulations, faster – enabling us to deliver safer and more energy-efficient structures to our clients." - Jens Christian Bennetsen, Project Director, Ramboll

Customer:
Ramboll

Industry:
Construction / Architecture / Engineering

Deployment country:
Denmark

Solution:
Grid Computing, Technical Computing, Energy Efficiency, High Availability , Server Consolidation

IBM Business Partner:
Gridcore

Overview

With an annual turnover of over 174 million USD and employing more than 9,800 people, Ramboll is a multidisciplinary international engineering company. Founded in Copenhagen, Denmark in 1945, the company has since expanded its operations around the world, and now has over 200 offices in 19 countries – including the UAE, UK and USA.

Business need:
Engineering company Ramboll wanted to continue to grow its client base, but was rapidly running out of computing capacity to support the resource-hungry simulations that its clients’ projects demanded.

Solution:
Ramboll engaged IBM Business Partner Gridcore to deploy a high-performance computing solution comprising eight 16-core IBM® BladeCenter® HS23 servers and one IBM System x® 3650 class server.

Benefits:
Enables simulations to be completed up to seven times faster – allowing the company to support a larger client base and run more complex projects. Dense, consolidated footprint drives significant savings on operational IT costs.

Case Study

To read a Danish version of this case study, click here

With an annual turnover of over 174 million USD and employing more than 9,800 people, Ramboll is a multidisciplinary international engineering company. Founded in Copenhagen, Denmark in 1945, the company has since expanded its operations around the world, and now has over 200 offices in 19 countries – including the UAE, UK and USA.

Tackling barriers to expansion
With its business growing rapidly, Ramboll wanted to continue the trend by gaining the ability to process a larger number of its clients’ resource-hungry simulation workloads while continuing to meet tight deadlines.

As Jens Christian Bennetsen, Project Director at Ramboll explains: “My department focuses on computational fluid dynamics for many different engineering disciplines, and for two specific industries in particular – oil and gas, and construction. Many of our oil and gas projects require us to use digital modeling to understand the risks posed by fire hazards, and design counter-measures to mitigate them. This is extremely important, as it helps to ensure the safety of people working in dangerous environments.

“Similarly, for construction projects on large, tall buildings, we need to run simulations to predict the combined effect of thermal radiation from the sun, right down to the level of small thermal mannequins such as humans, IT equipment, and electric lighting. Using these insights, we can design highly energy-efficient spaces for our clients.”

To deliver the accurate insights needed to inform the design of safe, energy-efficient structures, Ramboll relies on a suite of fluid dynamics modeling software, including ANSYS, FDS and OpenFOAM. In the past, the company ran its simulations using eight-core workstations, which it clustered together ad hoc according to the demands of each processing task.

“Using clustered workstations to process our workloads presented a significant challenge as our client base expanded,” says Bennetsen. “We were only able to cluster three workstations before we suffered a massive drop in performance. For simulations requiring millimetric detail, the computational demands can be up to ten times as high as those at a lower resolution – and because we could only work with 24 cores, complex workloads took a long time to execute.

“With our client base growing, there was a strong business need to process a larger number of complex workloads within the same timeframe.”

In addition to the issue of scalability, Ramboll’s distributed physical hardware was complex and costly to maintain. To solve these challenges, the company decided to search for a specialized high-performance computing (HPC) solution.

Selecting a HPC solution
Ramboll issued a request for proposals, and reviewed HPC offerings from a number of different vendors. Ultimately, the company selected a HPC solution based on IBM BladeCenter technology from IBM Business Partner Gridcore.

“We were extremely impressed with the IBM BladeCenter solution Gridcore presented,” says Bennetsen. “We had previously booked time on Gridcore’s massive cluster in Sweden to process some especially demanding workloads, so we were familiar with the GUI as well as the team there. We knew that we needed a partner with HPC expertise to deliver our solution, and Gridcore fit the bill perfectly.”

Working together with an expert team from Gridcore, Ramboll deployed eight 16-core IBM BladeCenter HS23 blades running CentOS Linux and one IBM System x3650 class server with an NVIDIA Quadro FX 3800 GPU at its Copenhagen datacenter. Today, the company can utilize up to a total of 128 cores – each of which has twice the performance of a single core from the legacy solution – to process its most demanding simulations.

Running simulations at lightning speed
Thanks to its IBM BladeCenter solution, Ramboll has dramatically accelerated its simulation processing – enabling the company to meet the computational demands of a larger client base, and support its growth objective.

“Using our IBM BladeCenter HPC solution, we can complete some simulations up to seven times faster than we could with our legacy systems,” says Bennetsen.

“In addition to enabling us to process the additional workloads that come with a larger client base, our HPC solution lets us perform multiple simulations for individual clients in the same amount of time that it used to take us to complete a single simulation on the legacy system. The more simulations we can run, the more accurate insights we can deliver for our clients, so this is an aspect of the solution that adds tremendous value to our offering.”

Because its compute power is consolidated into a single, dense footprint, the IBM solution has cut the IT maintenance costs associated with its distributed workstation environment.

Extending the solution
Following the success of the IBM HPC cluster, Ramboll is now expanding the solution to serve the needs of other departments.

“Our IBM cluster has become a key differentiator, as most of our competitors lack HPC capabilities,” says Bennetsen. “For this reason, there is now a lot of interest in the IBM solution from business users in different departments.

“For example, we have now added two 16-core HS23 blades to the IBM BladeCenter solution and an IBM System x3650 M4 server with an NVIDIA Quadro FX 4000 GPU, optimized for the finite element calculations performed by our structural analysis department. When other departments see the success we are having, we predict there will be even greater demand. In fact, we are already discussing the possibility of extending the solution to a further two departments.”

Bennetsen concludes: “We are extremely impressed by the positive impact that our IBM BladeCenter cluster is having on the business. Thanks to IBM and Gridcore, we can process more detailed simulations, faster – enabling us to deliver safer and more energy-efficient structures to our clients.”

Products and services used

IBM products and services that were used in this case study.

Hardware:
BladeCenter, BladeCenter HS23, BladeCenter running Linux - Other, System x, System x: System x3650 M3, System x: System x3650 M4

Operating system:
Linux

Legal Information

© Copyright IBM Corporation 2013 IBM Danmark ApS Nymoellevej 91 2800 Kgs. Lyngby Denmark Produced in Denmark April 2013 IBM, the IBM logo, ibm.com, BladeCenter and System x are trademarks of International Business Machines Corp., registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the web at: www.ibm.com/legal/copytrade.shtml. Linux is a registered trademark of Linus Torvalds in the United States, other countries, or both. This document is current as of the initial date of publication and may be changed by IBM at any time. Not all offerings are available in every country in which IBM operates. The client examples cited are presented for illustrative purposes only. Actual performance results may vary depending on specific configurations and operating conditions. It is the user’s responsibility to evaluate and verify the operation of any other products or programs with IBM products and programs. THE INFORMATION IN THIS DOCUMENT IS PROVIDED “AS IS” WITHOUT ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING WITHOUT ANY WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND ANY WARRANTY OR CONDITION OF NON-INFRINGEMENT. IBM products are warranted according to the terms and conditions of the agreements under which they are provided. The client is responsible for ensuring compliance with laws and regulations applicable to it. IBM does not provide legal advice or represent or warrant that its services or products will ensure that the client is in compliance with any law or regulation.