A complete picture of the processes occurring on seismically active faults is the essential prerequisite for the reliable assessment and mitigation of earthquake hazards. But earthquake parameters vary greatly over the fault in large earthquakes. For example, earthquake fault slip can range from 1-15 metres on different portions of the fault for any earthquake. Further slow slip has also been observed after a recent earthquake. In addition, earthquake damage depends directly on the speed at which the Earth's crust fractures. Our earlier studies showed that it is possible for this rupture velocity to be very high on some portions of the fault. Thus, sites close to regions of higher rupture velocity would be damaged more than other regions. Advance knowledge of such sites, based on earthquakes that have already occurred, is invaluable for making decisions on siting of critical structures (for example, dams, bridges, power plants, etc.). This proposal aims to quantify the distribution of slip, rupture speed, stress drops, and other fundamental source characteristics over the fault for several recent large earthquakes, by analysing seismograms. The results will also be used to increase our understanding of active tectonic areas. The relevant equations will be solved using a massively parallel computer cluster. The recent occurrence of several large earthquakes makes this investigation timely.