Hewlett Packard Enterprise Development LP

Joaquim Gomes Da Costa Eulalio De Souza

Martin Foltin

Eun Sub Lee

Alex Ferreira Ramires Trajano

Jefferson Rodrigo Alves Cavalcante

Glaucimar Da Silva Aguiar

Monycky Vasconcelos Frazao

Rodrigo Jose Da Rosa Antunes

Lucas Leite

Francisco Plínio Oliveira Silveira

Arthur Carvalho Walraven Da Cunha

Systems and methods are provided to improve traditional chip processing. Using crossbar computations, the convolution layer can be flattened into vectors, and the vectors can be grouped into a matrix where each row or column is a flattened filter. Each submatrix of the input corresponding to a position of a convolution window is also flattened into a vector. The convolution is computed as the dot product of each input vector and the filter matrix. Using intra-crossbar computations, the unused space of the crossbars is used to store replicas of the filters matrices and the unused space in XIN is used to store more elements of the input. In inter-crossbar computations, the unused crossbars are used to store replicas of the filters matrices and the unused XINs are used to store more elements of the input. Then, the method performs multiple convolution iterations in a single step.

System and method for processing convolutions on crossbar-based neural network accelerators for increased inference throughput

Growing community of inventors

Porto Alegre, Brazil

Joaquim Gomes Da Costa Eulalio De Souza