Parallel Algorithm Design and Performance Evaluation of FDTD on 3 Different Architectures: Cluster, Homogeneous Multicore and Cell/B.E.

Microwave tomography for breast cancer detection on Cell broadband engine processors

“…In [36,37], we parallelized FDTD on Cell BE and compared the results with a shared memory architecture and a cluster of distributed memory multicores. Cell BE achieved a speedup of 14.14 over AMD Athlon and 7.05 over AMD Opteron at the processor chip level.…”

Acceleration of FDTD Method Using a Novel Algorithm on the Cell B.E.

“…However, the analysis region size has remained limited by the local store capacity. M. Xu et al [8], [9] implemented the two-dimensional FDTD method on the IBM Blade QS20 (3.2 GHz) and reported computation speed that is 14.14 times that of an Opteron (AMD Inc.) processor and 7.05 times that of an Opteron processor. The analysis region was allocated to the main memory, which was used as shared memory.…”

Implementation and Optimization of cGA+LS to solve Capacitated VRP over Cell/B.E.

“…Cell BE processor is architected for multimedia applications with regular processing requirements. However, Cell BE shows significant speedups even for the problems with nonuniform memory access [14,15]. GAs are well known for their large memory requirements and non-uniform memory access patterns.…”