A patent filing made by IBM has detailed the design of world’s first 100 PetaFlops supercomputer.
The patent application, which is more than 600 pages long, details the architecture of the BlueGene/Q “Sequoia” supercomputer system, which the Big Blue is building for the Lawrence Livermore National Laboratory, which is capable of reaching 20 PetaFlops and an updated system which will be capable of reaching 107 PetaFlops.
The 20 Petaflops supercomputer will be based on 1.6 million A2 processing cores in 96 racks. The company claimed that each node will be capable of consuming around 30 watts of power meaning that the supercomputer will be able to reach 107 PetaFlops using just 15.7 MW of power. The supercomputer is expected to go online in 2012.
“A Multi-Petascale Highly Efficient Parallel Supercomputer of 100 petaOPS-scale computing, at decreased cost, power and footprint, and that allows for a maximum packaging density of processing nodes from an interconnect point of view,” the IBM patent application explained.
“The Supercomputer exploits technological advances in VLSI that enables a computing model where many processors can be integrated into a single Application Specific Integrated Circuit (ASIC),” it added.