Abstract
The multi-mesh of trees (MMT) is an efficient interconnection network, which is a hybrid product of multi-mesh and mesh of trees networks. Several parallel algorithms have been efficiently mapped on this architecture for various problems including root finding, langrage interpolation, matrix multiplication and parallel sorting. In this paper, we propose a parallel algorithm for prefix computation for N = n <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">4</sup> data on an iV-processor multi-mesh of trees which takes 3.75 log N + 6 time for N data points. This can be compared with parallel prefix on the similar architecture, i.e., multi-mesh (MM) which requires 13N <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1/4</sup> + log N -1 time.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call