US 11,810,223 B2
Controlling multi-GPU execution of kernels by kernel portion and resource region based dependencies
Florian A. Strauss, Los Angeles, CA (US); and Mark Evan Cerny, Los Angeles, CA (US)
Assigned to Sony Interactive Entertainment LLC, San Mateo, CA (US)
Filed by Sony Interactive Entertainment LLC, San Mateo, CA (US)
Filed on Mar. 28, 2022, as Appl. No. 17/706,558.
Application 17/706,558 is a continuation of application No. 16/861,049, filed on Apr. 28, 2020, granted, now 11,288,765.
Prior Publication US 2022/0222768 A1, Jul. 14, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. G06T 1/20 (2006.01); G06F 9/48 (2006.01); G06T 1/60 (2006.01); G06T 15/00 (2011.01)
CPC G06T 1/20 (2013.01) [G06F 9/4881 (2013.01); G06T 1/60 (2013.01); G06T 15/005 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method, comprising:
dividing a first kernel into a plurality of portions during execution of an application;
assigning a first portion of the first kernel to a first graphics processing unit (GPU) for processing;
assigning a second portion of the first kernel to a second GPU for processing;
executing to completion the first portion of the first kernel by the first GPU;
storing first data from the executing to completion the first portion of the first kernel on a first storage resource;
generating dependency data after the storing the first data on the first storage resource; and
executing a second kernel after determining that the dependency data has been generated,
wherein the executing the second kernel begins before the second portion of the first kernel has completed processing by the second GPU.