- Serial algorithms to evaluate the Gabor transform of a discrete signal are bound by the length of signal for which the transform can be evaluated. The time taken, if machine memory and other factors are ignored, grows as order O(N2 ) making it unsuitable for transforms of large signal lengths. In this paper, we present a parallel algorithm to generate the computationally intensive Gabor transforms of a discrete signal. This algorithm is suited to both shared memory and message passing systems and is independent of the underlying hardware. The performance was tested on a Linux cluster build on the message passing mode land linear speedup was observed. The network overheads are measured and the algorithm performance is proven to be independent of network contention for practical cases, although it provides a theoretical bound on achievable speedup. The design of the algorithm parallelizes the most computationally intensive subprocess of transform evaluation and also makes it suitable f...