Increasingly computational resources are required in mathematical modeling, then interpreted programming languages gain more space in scientific computing and applied mathematics.The purpose of this study was to evaluate the structures of the languages interpreted MatLab and Octave, important in the writing of codes to obtain approximate solutions of EDP's. Euler's equations that describing the flow of ideal fluids were used and their solutions were found via finite volume numerical methods: Lax-Friedrich, Nessyahu-Tadmor, Lax-Wendroff and Godunov. In all situations, we compared MatLab and Octave with the compiled language Fortran 90. The results showed that execution time and computational cost depend not only on the method chosen to solve a given mathematical problem, but also on the programming language chosen to implement the methods, coupled with the programming technique used.