Parallel computing with load balancing on heterogeneous distributed systems

Object-oriented programming of distributed iterative equation solvers

“…It should also be noted that one of the advantages of distributed computing is the increase in available memory, hence reducing the need for disk storage of data during calculation phases [13]. A further feature of PC networks is their heterogeneous nature, this further complicates the load balancing issue, and work has recently been reported in this area [14,15].…”

Object oriented implementation of distributed finite element analysis in .NET

“…It should also be noted that one of the advantages of distributed computing is the increase in available memory, hence reducing the need for disk storage of data during calculation phases [14]. A further feature of PC networks is their heterogeneous nature, this further complicates the load balancing issue, and work has recently been reported in this area [15,16].…”

Two-level parallelization of a fluid mechanics algorithm exploiting hardware heterogeneity

“…Several different approaches exist and it is difficult -as experienced by the authors -to assess the potential of each in order to decide about future algorithmic developments. Homogeneous CPU clusters are nowadays well understood, as are heterogeneous CPU clusters [29]. While GPGPU-computing added a new layer of complexity and complications to HPC, even complex algorithms are now running on clusters of such hardware, sometimes enabling tremendous speedups over the traditional CPU clusters while using similar programming techniques [22,21].…”