Abstract
We show a simple parallel acceleration (from about 2 n to about 1.4 n log nparallel arithmetic steps) of the straightforward parallelization of the substitution algorithm for a nonsingular triangular linear system of n equations. This only requires that we increase by less than 3 times the overall number of flops (or the potential work) of the former algorithm. The previous parallel acceleration of the substitution algorithm in [1] was slower than ours by the factor log n.
Full Text
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