US 12,265,915 B2
Composable neural network kernels
Chao Liu, Austin, TX (US); Daniel Isamu Lowell, Austin, TX (US); Wen Heng Chung, Austin, TX (US); and Jing Zhang, Austin, TX (US)
Assigned to Advanced Micro Devices, Inc., Santa Clara, CA (US)
Filed by Advanced Micro Devices, Inc., Santa Clara, CA (US)
Filed on Dec. 30, 2020, as Appl. No. 17/138,709.
Application 17/138,709 is a continuation in part of application No. 16/779,557, filed on Jan. 31, 2020.
Claims priority of provisional application 63/024,376, filed on May 13, 2020.
Claims priority of provisional application 62/927,603, filed on Oct. 29, 2019.
Claims priority of provisional application 62/925,168, filed on Oct. 23, 2019.
Claims priority of provisional application 62/867,766, filed on Jun. 27, 2019.
Prior Publication US 2021/0117806 A1, Apr. 22, 2021
This patent is subject to a terminal disclaimer.
Int. Cl. G06N 3/10 (2006.01); G06F 8/41 (2018.01); G06F 17/16 (2006.01)
CPC G06N 3/10 (2013.01) [G06F 8/41 (2013.01); G06F 17/16 (2013.01)] 22 Claims
OG exemplary drawing
 
1. A method for manipulating a generic tensor, the method comprising:
receiving a request to perform an operation on generic tensor raw data associated with a generic tensor; and
responsive to the request, performing the operation on generic tensor raw data, the performing the operation including mapping a tensor coordinate specified by the request to a memory address, the mapping including evaluating a delta function associated with a type of the request,
wherein the operation generates a modified generic tensor descriptor based on a generic tensor descriptor without modifying the generic tensor raw data,
wherein the delta function indicates a difference between a previously calculated memory address and the memory address, the difference determined based on a difference between the tensor coordinate and a previously calculated tensor coordinate for the previously calculated memory address.