US 12,147,851 B2
System, method and recording medium for temperature-aware task scheduling
I-Hsin Chung, Chappaqua, NY (US); Huan Hu, Yorktown Heights, NY (US); and Wei Tan, Elmsford, NY (US)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on May 25, 2023, as Appl. No. 18/201,929.
Application 18/201,929 is a continuation of application No. 17/319,217, filed on May 13, 2021, granted, now 11,740,945.
Application 17/319,217 is a continuation of application No. 16/777,263, filed on Jan. 30, 2020, granted, now 11,042,420, issued on Jun. 22, 2021.
Application 16/777,263 is a continuation of application No. 16/056,670, filed on Aug. 7, 2018, granted, now 10,621,012, issued on Apr. 14, 2020.
Application 16/056,670 is a continuation of application No. 15/238,258, filed on Aug. 16, 2016, granted, now 10,133,610, issued on Nov. 20, 2018.
Prior Publication US 2023/0297443 A1, Sep. 21, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 9/50 (2006.01); G06F 3/06 (2006.01); G06F 9/48 (2006.01); G06T 1/20 (2006.01)
CPC G06F 9/5094 (2013.01) [G06F 3/061 (2013.01); G06F 3/0653 (2013.01); G06F 3/0673 (2013.01); G06F 9/4893 (2013.01); G06F 9/5016 (2013.01); G06T 1/20 (2013.01); G06T 2200/28 (2013.01)] 3 Claims
OG exemplary drawing
 
1. A computer-implemented method for facilitating a task in a multi-graphical processing unit (GPU) environment, the method comprising:
learning acceptance conditions to accept a task to one or more GPUs in the multi-GPU environment, wherein the multi-GPU environment comprises two or more GPUs, based on a prior execution of the task on at least one of the one or more GPUs and based on a varying thermal characteristic of the at least one of the one or more GPUs during the prior execution, and wherein the acceptance conditions are variable based on time with respect to the varying thermal characteristic of an internal arithmetic logic unit (ALU) and a dynamic random-access memory (DRAM) of the at least one of the one or more GPUs at the variable times; and
executing the task to the at least one of the one or more GPUs based on the acceptance conditions of the at least one of the one or more GPUs as compared to each of the at least one of the one or more GPUs in the multi-GPU environment.