Advocates of Cloud Computing for HPC Workloads got a huge potential boost for their cause this week with the announcement that opening of the new UberCloud Marketplace.
The UberCloud started its free “HPC Experiment” back in 2012, which today has attracted more than 1000 organizations and individuals from HPC, computational fluid dynamics, finite element material analysis, multi-physics, chemistry, life sciences, biology, big data, and others. With the addition of the Ubercloud Marketplace, engineers and scientists can discover, try and buy the computing power and expertise on demand they need for their computational and data-intensive tasks.
With the limits of desktop workstations often unable to provide enough computing power, computer simulations taking too long, and the number of jobs too small to get quality results, engineers and scientists are looking for additional computing power beyond their desktop workstations. Now, the UberCloud Marketplace provides access to a wide variety of computing providers, software vendors, enabling tools, and independent experts to simplify and ease the search for the most suitable service providers and expertise, out of hundreds that joined UberCloud in the last 18 months.
How it Works
To get started, end users must first complete a “Request a Quote from Resource Providers.” They provide information about their application, software and licenses, network interconnect, main memory per node, number of parallel cores, total CPU usage, MIC/GPUs needed, storage, remote visualization, and instructions about timing, urgency, and location of resources. To help with this process, UberCloud’s site offers a LiveChat feature. Once your application is complted, the UberCloud takes care of the rest: automatically searching for suitable resource providers; collecting up to three quotes, and sending them to the end-user for consideration.
To date, the UberCloud has built 125 international teams exploring the end-to-end process of accessing and using remote computing resources, as a service, on demand; among them Team 118 with the end-user Rolls-Royce Germany about heat transfer in jet engines. “The outsourcing of computational workload to an external cluster allows the end user to distribute computing power in an efficient way; especially when the in-house computing resources are already at their limit,” said one of the end-users.
Resources for HPC as Service
As a comprehensive resource for HPC as a Service, the UberCloud also offers a services directory, case study discussion forums, technology and services webinars, and a monthly newsletter.
Read the Full Story.