This paper is part of a broader research work that intends to measure the impact that an urban form adapted to topography can have on mobility. Based on a literature review, a set of urban form variables with influence on car and pedestrian speed are defined. Considering the need, in such a complex context as cities, to isolate these variables from others, a GIS-based simulation is adopted, addressing large urban areas and calculating hundreds of thousands of routes simultaneously (using as many variables as possible). Salvador and San Francisco are used as opposing case studies, based on their respective networks of edges and nodes and grade data, available in an open database. Information regarding the number of buildings per edge is also added directly to the networks, to integrate speed and energy expenditure calculations through polynomial regressions. Two other variables, the intersections and the sinuosity of routes, are considered a posteriori in the calculation of routes through an origin-destination matrix based on those networks, using the Network Analyst extension of the ArcGIS software. The results of the calculation of minimizing pedestrian travel time, minimizing car travel time and minimizing car energy expenditure, are in line with previous results of this research and with the initial hypothesis: cars are penalized more than pedestrians in urban forms well adapted to topography. It is argued that this result is useful both for urban science and planning practice.