Cycle Computing Creates a 10,000 Core Cluster in the Cloud
Cycle Computing recently provisioned a 10,000-core, top 114-equivalent supercomputer utilizing its CycleCloud service. Since 2005, Cycle has helped clients maximize the world’s compute resources through its reliable, secure and elastic high performance computing (HPC) solutions, both internally and in the cloud.
CycleCloud massively scaled up client resources to perform hundreds of thousands of computational tests in a matter of eight hours. Once the results were produced, the customer could just "turn off" those resources with no further charges. Additionally, the 10,000-core cluster was run on a cost of only $1,060/hour.
Leveraging CycleCloud and Cycle’s HPC proficiency delivered these stats:
- Infrastructure: 1250 instances with 8-core / 7-GB RAM
- Cluster Size: 10,000 cores, 8.75 TB RAM, 2 PB of disk space total
- Scale: Comparable to #114 of Top 500 Supercomputer list
- Security: Engineered with HTTPS & 128/256-bit AES encryption
The end-user experience for using CycleCloud is:
- User Effort: Single click to start the cluster
- Start-up Time: Thousands of cores in minutes, full cluster in 45-minutes
- Up-front Capital Investment/Licensing Fees: $0
- Total CycleCloud and Infrastructure Cost: $1,060/hour
"Running HPC applications in the Cloud is difficult. Sometimes special hardware is required — such as GPUs or Infiniband networks," said John Barr, Research Director, The 451 Group. "While a large class of HPC applications can run effectively both on large, low cost, commodity clusters and in the Cloud. However, the idea of managing many thousands of jobs that collaborate to deliver insight is beyond many scientists who would prefer to focus on their area of specialization rather than becoming compute cluster or Cloud experts."
The CycleServer cluster management and analytics software was used to track utilization, diagnose performance and manage the progress of the scientific workflow. In total, 9.1 years of computing were completed in eight hours using server instances. The average cost of deploying a cluster of this size in-house is estimated in the millions of dollars, given the additional expense for hardware, software, networking, storage, personnel time and other factors.
"CycleCloud exemplifies the far-reaching capabilities of cloud computing to increase the speed of research using cost-effective compute power," said Jason Stowe, founder and CEO Cycle Computing. "Our ability to spin up large clusters with no user effort provides an agile option for researchers in life science, financial, engineering, insurance and other industries. With this repeatable 10,000 core cluster under our belts, our team is already working on the next generation of secure, mega-elastic, and fully supported cloud clusters that are both timeframe and bottom-line friendly."
To learn more about the development of the 10,000 core cluster and Cycle’s projects leading up to this accomplishment, please visit the Cycle Computing blog: Compute Cycles (http://blog.cyclecomputing.com/).
About Cycle Computing:
Cycle Computing, a bootstrapped, profitable software company, delivers proven, secure and flexible high performance computing (HPC) and data solutions since 2005. Cycle helps clients maximize existing infrastructure and speed computations on desktops, servers, and on-demand in the cloud. Thanks to our CycleServer HPC management software and our CycleCloud fully-supported & secured HPC clusters, Cycle clients experience faster time-to-market, decreased operating costs, and unprecedented service & support. Starting with three initial Fortune 100 clients, Cycle has grown to deploy proven implementations at Fortune 500s, SMBs and government and academic institutions including JP Morgan Chase, Purdue University, Pfizer and Lockheed Martin.