Back to results
US 12,657,003 B2Granted
Neural network engine with associated memory array
AssigneeAnaflash Inc.
InventorsSeung-Hwan Song, Sihwan Kim
Life of this patent
Estimating the term…
Abstract
A neural network engine configured to perform neural network computations. The neural network engine includes a first memory block configured to provide a first operand; a second memory block configured to provide a second operand; a synapse array configured to produce partial products of the first and second operands, wherein the synapse array is located outside of the first and second memory blocks; and, an accumulator array to perform summation of the partial products produced by the synapse array.
Claims19
A neural network engine to perform computations, comprising:
- a first memory block configured to provide a first operand;
- a second memory block configured to provide a second operand;
- a synapse array physically distinct from the first memory block and second memory block comprising a plurality of synapse circuit, wherein each synapse circuit is configured to:
- receive first complementary input signals representing the first operand;
- produce a partial product based on the first operand and the second operand;
- and
- output complementary partial product signals corresponding to the partial product;
- and,
- one or more pairs of differential bit lines, wherein each pair of the one or more differential bit lines is configured to carry the complementary partial product signals from each corresponding synapse circuit;
- and
- an accumulator array coupled to the synapse array via the one or more pairs of differential bit lines, the accumulator configured to perform summation of the complementary partial products signals via the one or more pairs of differential bit lines.
8 more elements
Dependencies are inferred from the claim text.