This thesis deals with the high performance computation problems and more specifically with those of scientific parallel computation for irregular real-world applications. In the first part, we describe a method for overlapping communications on parallel computers with distributed memory. This method has resulted in a generic computation scheme for the optimal packet size. We also tackle the problem of finding the optimal computation grain for the Cholesky factorization algorithm for dense matrices. The goal of this study is to exploit the irregularity induced by the matrix symmetry. Based on this work we have developped a portable software library providing an efficient application context for these techniques. The second part of this thesis presents and analyses a general algorithm for the computation of an efficient static scheduling of block computations, developped especially for a parallel direct sparse linear factorization based on a combination of 1D and 2D block distributions. Our solver uses a supernodal Fan-In approach and is fully driven by our static scheduling algorithm. Compared to the existing parallel direct solvers our solver shows very favorable performance results.