Abstract: Fast and accurate precipitation estimation is an important element of remote atmosphere monitoring, as it allows, for example, to correct short-term weather forecasts and the prediction of several types of meteorological threats. The paper presents methodology for calibrating precipitation estimation algorithm based on MSG SEVIRI sensor data, and Optimal Cloud Analysis product available via EumetCast transmission. Calibration is performed in a predefined coastal zone area, and utilizes the parallelized gradient computing method. In order to perform and validate results of estimation, reference weather radar data was provided by the Meteorology and Hydrology Research Institute (pol. Instytut Meteorologii i Gospodarki Wodnej - IMGW). The research was conducted using the Tryton supercomputer - the high performance computing environment of Gdansk University of Technology.