A review is presented on some attempts to understand turbulence by using high-performance computing (HPC). Among them is the study of turbulence by the direct numerical simulation (DNS) of incompressible turbulence in a periodic box with the number of grid points up to 4096(3). The DNS data shed some light on the Reynolds number (Re)-dependence of the statistics of turbulence at high Re and on the intermittency of energy dissipation. HPC is shown to be useful also for solving a multidimensional singular partial differential equation that determines the anomalous scaling of a passive scalar field advected by a random velocity field with an infinitely small correlation time.