Innovation & Quality


MEET THE TEAM

End User &Team Expert – Justin M. Morgan, PE Principal, Ocean Engineering & Analysis, Glosten Inc.

Software Provider – Aji Purwanto, Business Development Director, NUMECA International S.A.

Resource Provider – Richard Metzler, Software Engineer, CPU 24/7 GmbH

Technology Experts – Hilal Zitouni Korkut and Fethican Coskuner, UberCloud Inc.

 

USE CASE

Glosten  is  a  Seattle-based  consulting  engineering  firm  comprised  of  naval  architects,  marine engineers, electrical engineers and ocean engineers. Since the firm's inception in 1958, Glosten has repeatedly delivered innovative solutions to challenging problems in the marine industry. Design experience includes tugs, barges, research vessels, cruise vessels, passenger/car ferries, and special- purpose platforms.

In this project, we calculated the barehull resistance of the KRISO containership (KCS) in the cloud. The KRISO containership is a standard hull form frequently used as a benchmark case for computational fluid dynamics in the marine industry. Both basic hull form parameters and experimental results are available in published literature. Detailed information for the KCS test case is also available on the internet here:

https://www.nmri.go.jp/institutes/fluid_performance_evaluation/cfd_rd/cfdws05/

The purposes of this project were to become familiar with the mechanics of running a FINE™/Marine simulation in an UberCloud container and to assess the performance of the available hardware compared to resources currently used by the end-user.  The benchmark case was analyzed on local hardware, on virtual instances in the cloud, and on the bare-metal cloud solution offered by CPU 24/7 and UberCloud.  All simulations were run using version 4 of Numeca’s FINE™/Marine software.

The cloud resource provider CPU 24/7 GmbH is a leading provider of CAE as a Service solutions for all application areas of industrial and academic /university research and development. Headquartered in Potsdam/Germany, CPU 24/7 develops and operates unique on demand services for High Performance Computing that are based on the latest globally accepted industry standards for hardware, software, and applications.

 

CHALLENGES

No  challenges  were  experienced  in  downloading  project  files  into  the  FINE™/Marine  container, running the  simulation,  or  retrieving data. The  remote desktop user  interface  was  responsive without any significant delays. Logging into the system is simple and the Numeca software pre-installed in an UberCloud container runs without any user setup. The only user setup required is to adjust the display resolution.

 

PROCESS AND BENCHMARK RESULTS

The simulation was setup as a steady state solution, fixed in trim and heave to duplicate the conditions of the experimental data.  The half model mesh contains 1.6 million cells.  Simulation control variables in FINE™/Marine were as follows:

  • 300 time steps
  • Uniform time step = 5 sub-cycles
  • 8 non-linear iterations

The solution converges to a steady state resistance force within about 150 time steps; however, the simulation was allowed to run to completion on all platforms to provide a performance comparison.

The calculated total resistance coefficient for this model is 0.003574 compared to the experimental result of 0.00356, a 0.4% difference. Figure 1 illustrates the calculated wave field (top) compared to measured data (bottom).

Figure 1: Comparison of experimental and calculated results

The processors offered by CPU 24/7 and available through the UberCloud container provide a significant improvement in performance over local Glosten hardware and over virtual instances available through Amazon Web Services (AWS).  The AWS compute instance used here is the third generation, c3.8xlarge.  A fourth generation compute instance is available on AWS with Intel Xeon CPU E5-2666 v3; however, setting up a new virtual instance was considered too costly for this project.

Platform

Processor

FINE™/Marine

#cores

Time [hrs]

local

Intel Xeon CPU E5645 @ 2.4 GHz x 2

v4.2

12

6.9

CPU 24/7

Intel Xeon CPU E5-2690 v3 @ 2.6 GHz x 24

v4.1

12

3.0

CPU 24/7

Intel Xeon CPU E5-2690 v3 @ 2.6 GHz x 24

v4.1

16

2.5

CPU 24/7

Intel Xeon CPU E5-2690 v3 @ 2.6 GHz x 24

v4.1

24

1.7

AWS

Intel Xeon CPU E5-2680 v2 @ 2.8 GHz x 23

v4.2

16

3.5

AWS

Intel Xeon CPU E5-2680 v2 @ 2.8 GHz x 23

v4.2

24

2.9

Figure 2: Performance comparison. The difference between v4.2 and v4.1 is only in patches not affecting performance.

 

BENEFITS

This use case helped us understand the performance benefits offered by UberCloud and CPU 24/7.

Glosten considers the UberCloud service to be a viable alternative to a local server upgrade. Additional benefits include the on-demand access and use of the software and hardware resources, a reduction in overhead required to manage virtual instances and to maintain software updates.

 

CONCLUSIONS

  • We  showed  that  the  CPU  24/7  HPC  bare-metal  cloud  solution  provides  performance advantages for Numeca FINE™/Marine users who want to obtain higher throughput or analyze larger, more complex models.
  • CPU 24/7 and UberCloud effectively eliminate the need to maintain in-house HPC expertise.
  • The  container  approach  provides  immediate  access  to  high  performance  clusters  and application software without software or hardware setup delays.
  • The browser-based user interface is simple, robust, and responsive.

 

APPENDIX: UberCloud Application Containers for Numeca FINE™/Marine

UberCloud Containers are ready-to-execute packages of software. These packages are designed to deliver the tools that an engineer needs to complete his task in hand. In this experiment, the FINE™/Marine software has been pre-installed, configured, and tested, and were running on bare metal, without loss of performance. The software was ready to execute literally in an instant with no need to install software, deal with complex OS commands, or configure.

The UberCloud Container technology allows wide variety and selection for the engineers because the containers are portable from server to server, Cloud to Cloud. The Cloud operators or IT departments no longer need to limit the variety, since they no longer have to install, tune and maintain the underlying software. They can rely on the UberCloud Containers to cut through this complexity.

This technology also provides hardware abstraction, where the container is not tightly coupled with the server (the container and the software inside isn’t installed on the server in the traditional sense). Abstraction between the hardware and software stacks provides the ease of use and agility that bare metal environments lack.

 

Share this article

Comments

Leave first comment here...

LEAVE A COMMENT

 
  • Would you like to know more?
    Click below to get in touch.