“School of Computer Science”

Back to Papers Home
Back to Papers of School of Computer Science

Paper   IPM / Computer Science / 10984
School of Computer Science
  Title:   Parallel Lagrange interpolation on k-ary n-cubes with maximum channel utilization
  Author(s): 
1.  A. Mahabadi
2.  H. Sarbazi-Azad
3.  E. Khodaei
4.  K. Navi
  Status:   Published
  Journal: The Journal of Supercomputing
  No.:  1
  Vol.:  46
  Year:  2008
  Pages:   1-14
  Publisher(s):   Kluwer Academic Publishers
  Supported by:  IPM
  Abstract:
This paper proposes an efficient parallel algorithm for computing Lagrange interpolation on k-ary n-cube networks. This is done using the fact that a k-ary n-cube can be decomposed into n link-disjoint Hamiltonian cycles. Using these n link-disjoint cycles, we interpolate Lagrange polynomial using full bandwidth of the employed network. Communication in the main phase of the algorithm is based on an all-to-all broadcast algorithm on the n link-disjoint Hamiltonian cycles exploiting all network channels, and thus, resulting in high-efficiency in using network resources. A performance evaluation of the proposed algorithm reveals an optimum speedup for a typical range of system parameters used in current state-of-the-art implementations.

Download TeX format
back to top
scroll left or right