The UberCloud Experiment (also known as the HPC Experiment) started in July last year to explore the process of accessing and using remote HPC resources or HPC-as-a-Service. From an initial pool of 160 participating organizations located around the world, 25 teams were created, consisting of an industry end-user and their application, the software provider, the computational resource provider, and the HPC expert who handles the porting of the application onto the resource and serves as team manager.
Round 2, which began in December and features both CAE and life sciences applications, has attracted 300 participating organizations, with 20 established teams and more to come. The project management tool BaseCamp guides teams through 22 well-defined steps of the end-to-end process, while a new services directory, UberCloud Exhibit, lists UberCloud hardware, software and expertise services.
Now it is time to report back to the HPC community what we have learned so far, for example which industry applications have been implemented on remote computing resources and in the cloud; how the teams have faced and resolved major roadblocks; what the optimal end-to-end process looks like; and additional guidance and recommendations. While the Experiment continues, we are beginning to get invitations to present these findings at conferences. Here are some of them – we invite interested parties to stop by and talk to us:
If you would like to participate in Round 3 of the UberCloud HPC Experiment starting in April you can register now and we’ll send you additional information. And if you want to learn more about some of the major services used in the Experiment, please go to the interactive UberCloud Exhibit.