Modeling turbulent transport is a major goal in order to predict confinement issues in a tokamak plasma. The gyrokinetic framework considers a computational domain in five dimensions to look at kinetic issues in a plasma. Gyrokinetic simulations lead to huge computational needs. Up to now, the gyrokinetic code GYSELA performed large simulations using a few thousands of cores. The work proposed here improves GYSELA onto two points: memory scalability and execution time. The new solution allows the GYSELA code to scale well up to 64k cores.