How Cloud Computing Can Accelerate Your Projects
As your projects grow in complexity, so do your needs for higher end hardware. This is especially true for those using design software and even more so for those utilizing simulation software. When looking into purchasing new hardware, the costs will add up quick. Another thing to keep in mind is that with the rapid development of software and hardware trends, new hardware should be selected with the idea of lasting 3-5 years. Cloud computing provides some great opportunity to reduce some of the burden on your local hardware resources, accelerating your time to market and freeing you up to continue working on other projects.
Topics covered in this post:
How is Your Hardware Utilized?
The intended use for your computing hardware is an essential determination while looking into what to purchase. For example, you may think that purchasing a computer with as many cores as possible will be the go-to move, however, this largely depends on the intended software’s capabilities. For example, when carrying out design with SOLIDWORKS, many of the processes are single threaded, meaning that having more cores will not significantly improve your performance. In this case, it’s better to go after a higher processor clock speed.
When it comes to simulation, we have to look at things a bit differently. Software such as SOLIDWORKS Simulation or SIMULIA have the capability to utilize multiple cores, which will accelerate certain aspects of performance such as solve times. You can determine how many cores you would like to utilize, which will then tie them up for the duration of the simulation.
What is HPC?
HPC stands for High Performance Computing. In the past, many companies that frequently completed computationally expensive work, such as simulation, would often install HPC hardware on-premise. This allowed the simulations to be off-loaded to the HPC hardware and allowed the users to continue working on their local computers. This increased the efficiency of work, but just as mentioned earlier, this hardware still had a life expectancy of approximately 3-5 years. It’s important to note that this 3-5 years is referencing the relevancy of the hardware, not necessarily the actual life-span.
This 3-5 years hurts even more when you find out the cost of HPC hardware and additional costs. For example, you could expect to pay approximately $15,000 for a dual CPU, 32 core, 64 GB blade to be mounted in a server room, which breaks down to $3,000 per year if you stretch it to the far end of its life expectancy. To make things worse, hardware costs only contribute about 14% to the total infrastructure cost.
Check out this whitepaper if you want to take a more in depth look at this data!
Cloud Solutions
Having HPC hardware installed on premise still makes sense for certain companies, but most simulation users aren’t able to justify that cost. Luckily, cloud computing has advanced rapidly in the recent past. Cloud computing allows you to off-load your computationally expensive work (i.e. simulations) to HPC hardware maintained by external companies for a fee.
While the rate per hour when using cloud solution may be higher than having HPC on-premise, it’s important to consider that users will not be utilizing the HPC hardware 100% of the time. To have a true comparison of value, users should divide the total cost by the expected lifespan and then multiply that by the expected rate of actual usage. This calculation can be done on a per core basis and compared directly to that of the rate of cloud computing.
Cloud Computing with SIMULIA
Just as with most simulation solutions, you can utilize your local hardware to carry out simulations with SIMULIA. The number of cores that you can utilize varies from role to role. In addition to locally, SIMULIA has built in cloud computing capabilities. With each role purchased, you are provided with a number of dedicated cloud computing cores. This will allow you to set up the simulation locally and run it via cloud computing, freeing up your computer to allow you to continue working locally at no additional cost. This number of dedicated cloud cores also varies from role to role.
It doesn’t end there. There are two more capabilities available for cloud computing with SIMULIA.
Tokens
Tokens are reusable add-ons that allow you to add additional computing power for your consistent workloads. Tokens are checked out as a simulation is running and checked back in when it is complete for others to use. This allows you to consistently have high end computing power without the need for sustaining your own hardware and maintaining the infrastructure costs.
Credits
In addition to Tokens, there are credits available for your peak demands. If you have a project that needs results YESTERDAY, then you can greatly accelerate your simulations for that out-of-ordinary peak demand with the use of credits. Credits are consumed per simulation and are not re-usable.
SIMULIA currently has capabilities to provide up to 144 cores per simulation. The Abaqus solver is relatively linear in relation to the number of cores being used to run the solve. So, for example, if you had a simulation that takes 4 hours with 8 cores, then with 16 cores, it would take approximately 2 hours!
NOTE: Even if using cloud computing, you will still want a certified workstation, as everything prior to the solve will still be utilizing your local hardware.
Summary
Cloud computing is being used more and more and provides excellent benefits for many users. High Performance Computing (HPC) capabilities have been around for a while, but were previously only available to companies with extensive resources that could justify the cost with a high utilization rate. The movement for HPC power being provided via cloud services brings the power to a much wider audience.
With SIMULIA, regardless if you’re just getting started in simulation or a seasoned user, you can have the HPC resources of an established enterprise at your fingertips without the expensive and time-consuming implementation of installing and initializing all of the hardware and infrastructure.
Contact us today to see how SIMULIA can benefit your company!
Matt is a Customer Success Manager at SWYFT Solutions and started using SOLIDWORKS while earning his Bachelor’s in Mechanical Engineering at Grand Valley State University and has been hooked ever since! He now specializes in 3DEXPERIENCE Works solutions, including 3DEXPERIENCE SOLIDWORKS. Outside of work, Matt loves to travel, hike, and discover new music.
Connect with Matt on LinkedIn