Varian, Inc. is a leading Scientific Instruments Company, with a reputation for high quality products for scientific industries. Researchers run calculation-intensive Monte Carlo simulations of future products using partial differential equation mathematical models.
Varian needed to simulate a design for a mass spectrometer, a very compute intensive operation. The simulations required several thousand compute hours, and nearly 6 calendar weeks on an internal pool of processors. With a conference deadline for completed results looming, they needed to get results done faster.
Varian worked with Cycle Computing, an Advanced Technology Partner in the AWS Partner Network (APN), who developed CycleCloud to address the growing need to rapidly deploy clusters for pay-as-you-go intensive computing needs. Cycle Computing supported Varian’s implementation of CycleCloud for scientific computing needs in the cloud.
By provisioning a fully secured cluster that grew to several hundred CPUs runnng on Amazon Elastic Compute Cloud (Amazon EC2), CycleCloud dynamically scaled up to execute the simulation, then shut down when calculations completed. Rather than purchasing hardware, Varian was able to run a several-week calculation in under a day using CycleCloud and AWS.
“With AWS and CycleCloud, we were able to access a supercomputer with several hundred cores for a day, obtaining our modeling results just in time for the conference presentation and avoiding the overhead of buying and managing our own computers,” says Ken Newton, R&D Scientist Varian, Inc.
- Lower costs to operate: CycleCloud leverages the AWS pay-per-use model for fully secure, provisioned clusters, removing upfront capital expenditures and the costs of an operations team, while enabling access to tremendous compute power.
- Speed to market: Running a calculation on 1 machine for 100 hours on Amazon EC2 costs the same as using 100 machines for 1 hour. This empowers researchers to get results and develop products faster, while only using the servers they need.
In summary, Varian used AWS and CycleCloud to dramatically decrease the time required to compute product design simulations, while decreasing costs. By easily provisioning a secure cluster with the encryption required to execute sensitive calculations on the cloud, AWS makes high performance cloud computing easy.