FIELD: computing technology.
SUBSTANCE: invention relates to the field of computing technology using neural processors. The technical result of the claimed technical solution is achieved by the fact that provided in the claimed solution are a neural network processor and a compression unit configured to receive an uncompressed portion of data generated in the neural network processor, when compressing the data, each bit in part of the mask corresponds to a byte in the uncompressed portion of data, wherein each bit in part of the mask is set to logical zero, wherein the corresponding byte in the uncompressed portion of data is zero, and is set to a logical unit wherein the corresponding byte in the uncompressed portion of data is non-zero; part of the data of the compressed output portion of the data is therein generated and the compressed output portion of the data is output, wherein the compressed output portion of the data comprises part of the mask and part of the data.
EFFECT: increase in the productivity of data processing with simultaneous reduction in the energy consumption.
15 cl, 11 dwg
