In high-performance computing (HPC) systems, resources are allocated to users according to configurable scheduling policies. Unfortunately, these policies often create resource fragments which reduce overall resource utilization. In this paper, we propose an HPC architecture which maximizes the resource utilization by allocating unused resource fragments for jobs which are malleable in the sense that these jobs can be created and terminated at any time without severe penalties. Hence, the whole HPC system can utilize almost 100% of its computing resources by filling its schedule with these malleable jobs. Malleable job computation can then be offered at a lower price for potential customers, allowing for multi-pricing schemes for the HPC system provider.
IOS Press, Inc.
6751 Tepper Drive
Clifton, VA 20124
Tel.: +1 703 830 6300
Fax: +1 703 830 2300 firstname.lastname@example.org
(Corporate matters and books only) IOS Press c/o Accucoms US, Inc.
For North America Sales and Customer Service
West Point Commons
Lansdale PA 19446
Tel.: +1 866 855 8967
Fax: +1 215 660 5042 email@example.com