skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Dynamic wavefront creation for processing units using a hybrid compactor

Abstract

A method, a non-transitory computer readable medium, and a processor for repacking dynamic wavefronts during program code execution on a processing unit, each dynamic wavefront including multiple threads are presented. If a branch instruction is detected, a determination is made whether all wavefronts following a same control path in the program code have reached a compaction point, which is the branch instruction. If no branch instruction is detected in executing the program code, a determination is made whether all wavefronts following the same control path have reached a reconvergence point, which is a beginning of a program code segment to be executed by both a taken branch and a not taken branch from a previous branch instruction. The dynamic wavefronts are repacked with all threads that follow the same control path, if all wavefronts following the same control path have reached the branch instruction or the reconvergence point.

Inventors:
; ;
Publication Date:
Research Org.:
Advanced Micro Devices, Inc., Sunnyvale, CA (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
1422752
Patent Number(s):
9,898,287
Application Number:
14/682,971
Assignee:
Advanced Micro Devices, Inc. (Sunnyvale, CA)
DOE Contract Number:  
AC52-07NA27344
Resource Type:
Patent
Resource Relation:
Patent File Date: 2015 Apr 09
Country of Publication:
United States
Language:
English
Subject:
97 MATHEMATICS AND COMPUTING

Citation Formats

Puthoor, Sooraj, Beckmann, Bradford M., and Yudanov, Dmitri. Dynamic wavefront creation for processing units using a hybrid compactor. United States: N. p., 2018. Web.
Puthoor, Sooraj, Beckmann, Bradford M., & Yudanov, Dmitri. Dynamic wavefront creation for processing units using a hybrid compactor. United States.
Puthoor, Sooraj, Beckmann, Bradford M., and Yudanov, Dmitri. 2018. "Dynamic wavefront creation for processing units using a hybrid compactor". United States. https://www.osti.gov/servlets/purl/1422752.
@article{osti_1422752,
title = {Dynamic wavefront creation for processing units using a hybrid compactor},
author = {Puthoor, Sooraj and Beckmann, Bradford M. and Yudanov, Dmitri},
abstractNote = {A method, a non-transitory computer readable medium, and a processor for repacking dynamic wavefronts during program code execution on a processing unit, each dynamic wavefront including multiple threads are presented. If a branch instruction is detected, a determination is made whether all wavefronts following a same control path in the program code have reached a compaction point, which is the branch instruction. If no branch instruction is detected in executing the program code, a determination is made whether all wavefronts following the same control path have reached a reconvergence point, which is a beginning of a program code segment to be executed by both a taken branch and a not taken branch from a previous branch instruction. The dynamic wavefronts are repacked with all threads that follow the same control path, if all wavefronts following the same control path have reached the branch instruction or the reconvergence point.},
doi = {},
url = {https://www.osti.gov/biblio/1422752}, journal = {},
number = ,
volume = ,
place = {United States},
year = {Tue Feb 20 00:00:00 EST 2018},
month = {Tue Feb 20 00:00:00 EST 2018}
}

Works referenced in this record:

Creating and Managing Links to Deduplication Information
patent-application, January 2011


Dynamic warp subdivision for integrated branch and memory latency divergence tolerance
patent-application, September 2011


Techniques for Handling Divergent Threads in a Multi-Threaded Processing System
patent-application, March 2013


Compiler-Controlled Region Scheduling for SIMD Execution of Threads
patent-application, June 2014


Microarchitectural Design Space Exploration Using an Architecture-Centric Approach
conference, December 2007


Dynamic warp formation: Efficient MIMD control flow on SIMD graphics hardware
journal, June 2009


Thread block compaction for efficient SIMT control flow
conference, February 2011


SIMD divergence optimization through intra-warp compaction
journal, July 2013