公开(公告)号：US10367590B2

公开(公告)日：2019-07-30

申请号：US15272616

申请日：2016-09-22

Applicant: Hewlett Packard Enterprise Development LP

Inventor： David Kielpinski ,  Jason Pelc ,  Thomas Van Vaerenbergh ,  Nikolas Tezak ,  Gabriel Joe Mendoza ,  Raymond G. Beausoleil

IPC: H04B10/80 ,  G06E3/00

Abstract: Examples described herein relate to concurrently performing operations on optical signals. In an example, a method includes providing, to an optical circuit, a first plurality of signals having a first optical property and encoding a first vector. A second plurality of signals is provided to the circuit that encodes a second vector and has a second optical property that is different from the first optical property. A first attribute-dependent operation is performed on the first plurality of signals via the circuit to perform a first matrix multiplication operation on the first vector, and concurrently, a second attribute-dependent operation is performed on the second plurality of signals to perform a second matrix multiplication operation on the second vector. The first matrix multiplication operation and the second matrix multiplication operation are different based on the first optical property being different from the second optical property.

公开(公告)号：US20180081388A1

公开(公告)日：2018-03-22

申请号：US15272616

申请日：2016-09-22

Applicant: Hewlett Packard Enterprise Development LP

Inventor： David Kielpinski ,  Jason Pelc ,  Thomas Van Vaerenbergh ,  Nikolas Tezak ,  Gabriel Joe Mendoza ,  Raymond G. Beausoleil

IPC: G06E1/00 ,  H04B10/80

CPC classification number: H04B10/801 ,  G06E3/008

Abstract: Examples described herein relate to concurrently performing operations on optical signals. In an example, a method includes providing, to an optical circuit, a first plurality of signals having a first optical property and encoding a first vector. A second plurality of signals is provided to the circuit that encodes a second vector and has a second optical property that is different from the first optical property. A first attribute-dependent operation is performed on the first plurality of signals via the circuit to perform a first matrix multiplication operation on the first vector, and concurrently, a second attribute-dependent operation is performed on the second plurality of signals to perform a second matrix multiplication operation on the second vector. The first matrix multiplication operation and the second matrix multiplication operation are different based on the first optical property being different from the second optical property.