Traditional sequential computers are based on the model introduced by John von Neumann. Sequence of instructions operates on a sequence of data. There are at least two limiting factors influencing performance of the sequential computer: CPU execution rate of instructions and read/write access to the system memory.
IBM Roadrunner is a hybrid design with 12,960IBMPowerXCell8i CPUs and 6,480AMDOpterondual-core processors (Barker et al, 2008). Roadrunner is different from many contemporary supercomputers in that it is a hybrid system which uses two different processor architectures whereas the usual supercomputers use only one as using one processor architecture is generally easier to design and implement. As a supercomputer, the Roadrunner is considered an Opteron cluster with Cell accelerators, as each node consists of a Cell attached to an Opteron core and the Opterons to each other.
Due to the recent advances in search technologies coupled with internet searches, computing has risen to higher levels of scales involving huge data sets. Organizations like Google and other competitors have developed a new concept of supercomputing called Data Intensive Super Computer (DISC) systems. Search engines use a single processor architecture using DISC systems. DISC systems differ from conventional supercomputers in their focus is on data: they acquire and maintain continually changing data sets, in addition to performing large-scale computations over the data. With the massive amounts of data arising from such diverse sources as telescope imagery, medical records, online transaction records, and web pages, DISC systems have the potential to achieve major advances in science, health care, business, and information access. DISC opens up many important research topics in system design, resource management, programming models, parallel algorithms, and applications.
IBM Roadrunner Architecture
It is already known that the Roadrunner uses two different processor architectures: Opteron and PowerXCell. Opterons are used both in the computational nodes feeding the Cells with useful data and in the system operations and communication nodes passing data between computing nodes and helping the operators running the system. Roadrunner has a total of 6912 Opteron processors (6480 computation, 432 operation), for a total of (12960+864) 13824 cores (Barker et al, 2008). The PowerXCell processors have one general purpose core (PPE), and eight special performance cores (SPE) forfloating pointoperations. Roadrunner has a total of 12960 PowerXCell processors, with 12960 PPE cores and 103680 SPE cores, for a total of 116640 cores.
DISC Systems Architecture
Google implements DISC systems in a cluster architecture where a group of CPUs are racked together and coupled to work in co-ordination. Google uses around 899 racks with each rack consisting of 88 dual CPU 2 GHz Intel Xeon servers with 2 GigaBytes of Random Access Memory (RAM) and 80 GigaBytes of Hard Disk space (Skrenta, 2004). The DISC system implemented by Google has 253 Teraflops of processing power. Google's supercomputer system is one of the most powerful rivaling even NASA or NSA.
IBM Roadrunner was built basically to handle 'Grand Challenge' problems. To address