DOE Patents title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Processing unit performance projection using dynamic hardware behaviors

Abstract

Methods for estimating accelerator performance for dynamic hardware behaviors are disclosed. Computer program code to be executed on a first processing unit is received, and an execution of the computer code on the first processing unit is monitored to determine a plurality of performance characteristics. A plurality of dynamic hardware behaviors is determined by applying a clustering algorithm to the performance characteristics, and an equivalent accelerator portion of computer code to be executed on a second processing unit is generated by translating a set of instructions in a first portion of computer code corresponding to a first one of the plurality of dynamic hardware behaviors to an equivalent set of instructions to be executed on the second processing unit. An estimated measure of performance for executing the equivalent accelerator portion on the second processing unit is determined for the first one of the plurality of dynamic hardware behaviors.

Inventors:
; ; ;
Issue Date:
Research Org.:
International Business Machines Corp., Armonk, NY (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
1632473
Patent Number(s):
10540737
Application Number:
15/852,450
Assignee:
International Business Machines Corporation (Armonk, NY)
DOE Contract Number:  
7216497
Resource Type:
Patent
Resource Relation:
Patent File Date: 12/22/2017
Country of Publication:
United States
Language:
English
Subject:
97 MATHEMATICS AND COMPUTING

Citation Formats

Artico, Fausto, Brunheroto, Jose R., Garcia, Juan Gonzalez, and Gonzalez, Nelson Mimura. Processing unit performance projection using dynamic hardware behaviors. United States: N. p., 2020. Web.
Artico, Fausto, Brunheroto, Jose R., Garcia, Juan Gonzalez, & Gonzalez, Nelson Mimura. Processing unit performance projection using dynamic hardware behaviors. United States.
Artico, Fausto, Brunheroto, Jose R., Garcia, Juan Gonzalez, and Gonzalez, Nelson Mimura. Tue . "Processing unit performance projection using dynamic hardware behaviors". United States. https://www.osti.gov/servlets/purl/1632473.
@article{osti_1632473,
title = {Processing unit performance projection using dynamic hardware behaviors},
author = {Artico, Fausto and Brunheroto, Jose R. and Garcia, Juan Gonzalez and Gonzalez, Nelson Mimura},
abstractNote = {Methods for estimating accelerator performance for dynamic hardware behaviors are disclosed. Computer program code to be executed on a first processing unit is received, and an execution of the computer code on the first processing unit is monitored to determine a plurality of performance characteristics. A plurality of dynamic hardware behaviors is determined by applying a clustering algorithm to the performance characteristics, and an equivalent accelerator portion of computer code to be executed on a second processing unit is generated by translating a set of instructions in a first portion of computer code corresponding to a first one of the plurality of dynamic hardware behaviors to an equivalent set of instructions to be executed on the second processing unit. An estimated measure of performance for executing the equivalent accelerator portion on the second processing unit is determined for the first one of the plurality of dynamic hardware behaviors.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {2020},
month = {1}
}

Works referenced in this record:

Energy-aware task consolidation on graphics processing unit (GPU)
patent, February 2014


System and method for providing automated computer language translation and verification
patent, October 2017