Scavenging Idle CPU Cycles for Creation of Inexpensive Supercomputing Power

Abstract

Grid computing is often regarded as a means for creating inexpensive “super-computing.” With grid computing organizations can optimize computing and data resources, pool such resources to support massive, compute intensive loads. This paper aims at demonstrating experiences gained during aggregation of large computational power acquired from idle workstations connected on the campus. The cycle harvesting is attained by integrating and configuring a suite of grid and middleware software technologies anchored by the Globus toolkit, further optimized to extract the large computational capacity in the form of MIPS and GFLOPS. NirmaGrid produced is coordination of available computational efforts by Scavenging of idle CPU cycles, which allows the sharing of a computational task among multiple processors.