Nan Jiang

US Patent:

59397128, Aug 17, 1999

Filed:

May 30, 1997

Appl. No.:

8/866033

Inventors:

Prasad V. Venugopal - Sunnyvale CA
Nan Jiang - Fremont CA

Assignee:

Hewlett-Packard Company - Palo Alto CA

International Classification:

G01D 534

US Classification:

25023113

Abstract:

The quasi-absolute encoder module determines absolute position within one mechanical rotation. The encoder module includes a three channel position encoder and detection circuitry that includes a read-only memory (ROM). There are three channels: A, B, and I. Channels A and B have four periods and channel B is offset from channel A by 90. degree. in phase. Channel I has a unique sequence. The address generated by sampling this unique sequence corresponds to an address in ROM. The address indicates the position of the rotating shaft on which the codewheel is attached.

US Patent:

59987847, Dec 7, 1999

Filed:

May 8, 1998

Appl. No.:

9/074990

Inventors:

Prasad V. Venugopal - Sunnyvale CA
Nan Jiang - Fremont CA

Assignee:

Hewlett-Packard Company - Palo Alto CA

International Classification:

G01D 534

US Classification:

25023113

Abstract:

The quasi-absolute encoder module determines absolute position within one mechanical rotation. The encoder module includes a three channel position encoder and detection circuitry that includes a read-only memory (ROM). There are three channels: A, B, and I. Channels A and B have four periods and channel B is offset from channel A by 90. degree. in phase. Channel I has a unique sequence. The address generated by sampling this unique sequence corresponds to an address in ROM. The address indicates the position of the rotating shaft on which the code wheel is attached.

US Patent:

20210036877, Feb 4, 2021

Filed:

Jul 24, 2020

Appl. No.:

16/938156

Inventors:

- Santa Clara CA, US
Nan Jiang - Santa Clara CA, US
Larry Robert Dennison - Menden MA, US
Gregory M. Thorson - Palo Alto CA, US

International Classification:

H04L 12/18
H04L 12/741

Abstract:

A network device configured to perform scalable, in-network computations is described. The network device is configured to process pull requests and/or push requests from a plurality of endpoints connected to the network. A collective communication primitive from a particular endpoint can be received at a network device. The collective communication primitive is associated with a multicast region of a shared global address space and is mapped to a plurality of participating endpoints. The network device is configured to perform an in-network computation based on information received from the participating endpoints before forwarding a response to the collective communication primitive back to one or more of the participating endpoints. The endpoints can inject pull requests (e.g., load commands) and/or push requests (e.g., store commands) into the network. A multicast capability enables tasks, such as a reduction operation, to be offloaded to hardware in the network device.

US Patent:

20210036881, Feb 4, 2021

Filed:

Jul 24, 2020

Appl. No.:

16/938044

Inventors:

- Santa Clara CA, US
Nan Jiang - Santa Clara CA, US
Larry Robert Dennison - Menden MA, US

International Classification:

H04L 12/18
G06F 9/50
H04L 12/927
H04L 12/813
H04L 12/801

Abstract:

A network device configured to perform scalable, in-network computations is described. The network device is configured to process pull requests and/or push requests from a plurality of endpoints connected to the network. A collective communication primitive from a particular endpoint can be received at a network device. The collective communication primitive is associated with a multicast region of a shared global address space and is mapped to a plurality of participating endpoints. The network device is configured to perform an in-network computation based on information received from the participating endpoints before forwarding a response to the collective communication primitive back to one or more of the participating endpoints. An injection policy comprising the issuing of credits enables each endpoint to limit the amount of collective communication primitives injected into the network simultaneously to reduce network congestion caused by increased network traffic due to the multicast capability of the network devices.

US Patent:

20210037107, Feb 4, 2021

Filed:

Jul 24, 2020

Appl. No.:

16/938097

Inventors:

- Santa Clara CA, US
Nan Jiang - Santa Clara CA, US
Larry Robert Dennison - Menden MA, US
Gregory M. Thorson - Palo Alto CA, US

International Classification:

H04L 29/08
H04L 12/18
H04L 12/741

Abstract:

A network device configured to perform scalable, in-network computations is described. The network device is configured to process pull requests and/or push requests from a plurality of endpoints connected to the network. A collective communication primitive from a particular endpoint can be received at a network device. The collective communication primitive is associated with a multicast region of a shared global address space and is mapped to a plurality of participating endpoints. The network device is configured to perform an in-network computation based on information received from the participating endpoints before forwarding a response to the collective communication primitive back to one or more of the participating endpoints. The endpoints can inject pull requests (e.g., load commands) and/or push requests (e.g., store commands) into the network. A multicast capability enables tasks, such as a reduction operation, to be offloaded to hardware in the network device.

US Patent:

20190297018, Sep 26, 2019

Filed:

Feb 15, 2019

Appl. No.:

16/277349

Inventors:

- Santa Clara CA, US
Nan Jiang - Santa Clara CA, US
John Wortman - Santa Clara CA, US
Alex Ishii - Santa Clara CA, US
Mark Hummel - Santa Clara CA, US
Rich Reeves - Santa Clara CA, US

International Classification:

H04L 12/801
H04L 12/825
H04L 12/26

Abstract:

Multiple processors are often used in computing systems to solve very large, complex problems, such as those encountered in artificial intelligence. Such processors typically exchange data among each other via an interconnect fabric (such as, e.g., a group of network connections and switches) in solving such complex problems. The amount of data injected into the interconnect fabric by the processors can at times overwhelm the interconnect fabric preventing some of the processors from communicating with each other. To address this problem, techniques are disclosed to enable, for example, processors that are connected to an interconnect fabric to coordinate and control the amount of data injected so that the interconnect fabric does not get overwhelmed.