Abstract: This study proposes a method for the detection and correction of errors in the implementation of the matrix multiplication supercomputers due to noise. The proposed method based on a comparison of the results of the two operations with using CUDA-technology. Operations are performed in parallel by the action of the same interference.