1 March, 2019
6SigmaRoom is a Computational Fluid Dynamics (CFD) data center simulation platform that provides increased levels of productivity for data center design, troubleshooting and operation.
The 6SigmaRoom solve process consists of several steps, from generating grid at the start, running CFD calculations and reassembling data at the end. Solve time is dependent on the complexity and the size of the model being solved.
To ensure speedy solves, 6SigmaRoom offers an array of solver options that account for businesses’ varying processing capabilities and modeling needs.
This feature comparison will help you decide which solver option is best for you.
Figure 1. 6SigmaRoom has accurate solving options to fit your modeling needs
Parallel Solving with 6SigmaRoom
6SigmaRoom has been developed to take full advantage of modern multi-core processors. Using multi-core (i.e. parallel) solving divides the computing workload using all physical cores on the processor to accelerate the solve.
The more cores you have at your disposal, the better the computing power, the faster the overall solve process. In other words, larger models will benefit from an increase in core count.
6SigmaRoom enables you to solve on your laptop or workstation or to submit the solve job to a central powerful solving machine.
Solving on an HPC Cluster
In general, a Cluster consists of hardware and a scheduler. The scheduler allows the user to decide where and how to solve their models.
Our Cluster Solver can manage large workflows with less latency by pushing information across computing nodes faster. If you are regularly solving large models, we suggest using our Cluster Scheduler to solve as it enables the user to run a single job over multiple nodes of the Cluster.
For a more detailed explanation of 6SigmaRoom’s Core Solving Power, please see our Solver Performance Handout here.
Solving on the Cloud
We also provide our users with the option to solve with Rescale, our platform partner for Cloud computing.
Figure 2. 6SigmaRoom also has the option to solve on the Cloud
The user can solve on a Cluster through the Cloud at an hourly rate. While the user will get the same results as our Cluster Scheduler, it’s a “pay-as-you-go” model, i.e. you “rent” a Cluster. The Cloud solving option allows users more flexibility in terms of hardware options.
Flexible Options to Fit Your Business Needs
Still not sure which is right for you? Ask yourself: Do you want to rent or buy?
If you already have access to a powerful computer, 6SigmaRoom’s parallel solver is likely the choice for you.
If you’re only solving occasionally (e.g. solving one large model annually) and/or do not have convenient access to a powerful computer, we suggest renting solve hours on our Cloud partner, Rescale, to save on costs. It is worth noting that renting hours on the Cloud can add up in the long-run.
If you’re solving numerous projects consistently, we suggest using our Cluster Scheduler. There are also several other compatible schedulers available, such as Microsoft HPC Pack on Windows and IBM Spectrum LSF on Linux. Investing in Cluster hardware may be an important long-term investment for your company’s goals.
Solver options are not always black and white. We suggest contacting us for a personalized consultation to learn how best to use our solving options to get the most out of your CFD analysis.
Blog written by: Danielle Gibson, Technical Marketing Writer
Other Recent Posts
The Digital Twin for Hyperscale Data Centers
In a world where artificial intelligence and machine learning are the new frontiers, cloud applicat…
CFD Simulation Should Be a Data Center Best Practice
Over the years that I have been involved in the data center industry, I have read numerous articles…