Methods and apparatus for matrix processing in a convolutional neural network
Described examples include an integrated circuit including a vector multiply unit including a plurality of multiply/accumulate nodes, in which the vector multiply unit is operable to provide an output from the multiply/accumulate nodes, a first data feeder operable to provide first data to the vector multiply unit in vector format, and a second data feeder operable to provide second data to the vector multiply unit in vector format.
1. A method comprising:
receiving, by a first data feeder, first data;
providing, by a first data feeder, a portion of the first data to a plurality of multiply/accumulate nodes of a vector multiply unit;
providing, by a second data feeder, second data to the plurality of multiply/accumulate nodes of the vector multiply unit;
rearranging, using the first data feeder and the second data feeder, the first data and the second data;
multiplying, by the multiply/accumulate nodes of the vector multiply unit, the first data and the second data as a partial product;
accumulating, by the multiply/accumulator nodes of the vector multiply unit, the multiplied first data and second data; and
receiving, from the multiply/accumulator nodes of the vector multiply unit by a feature processing unit, the accumulated multiplied first data and second data;
in which the rearranging is performed so that individual feature values in the accumulated multiplied first data and second data are generated by corresponding individual ones of the multiply/accumulator nodes.
2. The method of claim 1 in which the multiplying and accumulating provides an outer product of the first data and the second data in a clock cycle.
3. The method of claim 1 in which the first data is K feature planes and N values are read from the K feature planes in a clock cycle of the vector multiply unit.
4. The method of claim 3 in which the second data is multiple sets of weight data with L weight coefficients and output from the vector multiply unit is provided after K×L cycles of the vector multiply unit.
5. The method of claim 1 in which an output of the vector multiply unit is a convolution layer and is provided after a fixed number of cycles of the vector multiply unit.
6. The method of claim 1 in which the first data is one row of length N, in which the second data is one column of length M that are read in a first cycle and multiplied and accumulated in N×M nodes in the vector multiply unit in a second cycle and in which, after the plurality of iterations, the N×M nodes are a results matrix that is at least part of a complete results matrix.