A distributed approach is described for solving lineality (or linearity) space (LS) problems with large cardinalities and a large number of dimensions. The LS solution has applications in engineering, science, and business, and includes a subset of solutions of the more general extended linear complementarity problem (ELCP). A parallel MATLAB framework is employed and results are computed on an 8-node Rocks based cluster computer using Remote Procedure Calls (RPCs) and the MPICH2 Message Passing Interface (MPI). Results show that both approaches perform comparably when solving distributed LS problems. This indicates that when deciding which parallel approach to use, the implementation details particular to the method are the decisive factors, which in this investigation give MPICH2 MPI the advantage.

In this poster, an interactive offline power modeling and analysis tool based on the R programming language is described. Least squares regression coefficients are calculated with R scripts from Watt meter and /proc file system data collected on a 20 compute node high performance computing (HPC) cluster. The High Performance Computing Challenge (HPCC) benchmark suite is used to represent "real world" HPC computations. Ten /proc file system files are sampled for each benchmarked compute node. Results show power consumption of HPC applications can be modeled at the full cluster level using the /proc file system both inexpensively and conveniently providing insight into the links between HPC cluster applications and cluster power consumption.