The Monte Carlo platform GATE (Geant4 Application for Tomographic Emission) is now used for dosimetry applications such like radiotherapy and brachytherapy. The aim of the developments actually performed is to enable an application of the GATE platform in clinical routine. However, this perpective is only possible if the coputation time of the simulations is highly reduced. The new grid architectures, developed within the framework of EGEE European project, is there to answer this feauture. James pseudorandom number generator (RNG) is used in GATE simulations. Our work has consisted into parallelize the simulations on multiple processors part of the grid by partitioning a sequence of random numbers generated by the RNG into suitable independent subsequences. At this date, we produced a thousand of random status files of some KBs corresponding to each independent subsequences of the RNG. In order to facilitate the access to the grid for novice users or praticians, we developed all the functionalities to create, split, launch and monitor GATE simulations on the EGEE infrastructure on the web portal (GENIUS). The transparent and secure access to the grid is provided by the portal accessible from the following URL https://clrglop208.in2p3.fr. To launch simulations an authentification and an authorization are required. A demonstration concerning the usage of the portal will be shown. To highlight the speed-up obtain by launching simulations on the grid, we present some computing time tests. The results obtained show a gain factor which can go up to 20 that make us envisage a future usage of GATE simulations for treatment planning in clinical routine.