A dictionary based accelerated method of lossless compression and encryption of data with a small statistical redundancy. The sequence of the least distances obtained between every input numerical sample and the samples of the several dictionaries, produced by random number generators, and statistical compression is further applied. The process is accelerated utilizing a packed content addressable memory (CAM) for dictionary and input samples. CAM is scanned for a marking bit indicating a presence of a dictionary sample to find the least distance. CAM is packed and loaded with successful ranges of the amplitudes of both dictionary and input samples to reduce an address space and to accelerate the process of compression further. The least distances are founded for an unique alphabet of the input sequence, and the results are merged with an original input to accelerate the process of compression further.