Title :
Multicore Acceleration of CG Algorithms Using Blocked-Pipeline-Matching Techniques
Author :
Fernández, David M. ; Giannacopoulos, Dennis ; Gross, Warren J.
Author_Institution :
Dept. of Electr. & Comput. Eng., McGill Univ., Montreal, QC, Canada
Abstract :
To realize the acceleration potential of multicore computing environments computational electromagnetics researchers must address parallel programming paradigms early in application development. We present a new blocked-pipeline-matched sparse representation and show speedup results for the conjugate gradient method by parallelizing the sparse matrix-vector multiplication kernel on multicore systems for a set of finite element matrices to demonstrate the potential of this approach. Performance of up to 8.2 GFLOPS was obtained for the proposed vectorized format using four Intel-cores, 17 × more than the nonvectorized version.
Keywords :
acceleration measurement; conjugate gradient methods; finite element analysis; microprocessor chips; parallel programming; pipeline arithmetic; sparse matrices; vectors; GFLOPS; Intel-cores; acceleration potential; blocked-pipeline-matching techniques; computational electromagnetics researchers; conjugate gradient algorithms; finite element matrices; multicore acceleration; multicore computing environments; parallel programming paradigms; sparse matrix-vector multiplication kernel; sparse representation; Acceleration; Character generation; Computational electromagnetics; Concurrent computing; Finite element methods; Gradient methods; Kernel; Multicore processing; Parallel programming; Sparse matrices; Acceleration; blocked formats; conjugate gradient; multicore; sparse matrices; vector processor;
Journal_Title :
Magnetics, IEEE Transactions on
DOI :
10.1109/TMAG.2010.2044023
