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

Title: Managing the Performance/Error Tradeoff of Floating-point Intensive Applications

Journal Article · · ACM Transactions on Embedded Computing Systems
DOI:https://doi.org/10.1145/3126519· OSTI ID:1474343
 [1];  [2];  [3];  [4];  [1]
  1. Univ. of Waterloo, ON (Canada)
  2. James Madison Univ., Harrisonburg, VA (United States)
  3. Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
  4. McMaster Univ., Hamilton, ON (Canada)

Modern embedded systems are becoming more reliant on real-valued arithmetic as they employ mathematically complex vision algorithms and sensor signal processing. Double-precision floating point is the most commonly used precision in computer vision algorithm implementations. A single-precision floating point can provide a performance boost due to less memory transfers, less cache occupancy, and relatively faster mathematical operations on some architectures. However, adopting it can result in loss of accuracy. Identifying which parts of the program can run in single-precision floating point with low impact on error is a manual and tedious process. In this paper, we propose an automatic approach to identify parts of the program that have a low impact on error using shadow-value analysis. Our approach provides the user with a performance/error tradeoff, using which the user can decide how much accuracy can be sacrificed in return for performance improvement. We illustrate the impact of the approach using a well known implementation of Apriltag detection used in robotics vision. In conclusion, we demonstrate that an average 1.3x speedup can be achieved with no impact on tag detection, and a 1.7x speedup with only 4% false negatives.

Research Organization:
Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)
Sponsoring Organization:
USDOE National Nuclear Security Administration (NNSA)
Grant/Contract Number:
AC52-07NA27344
OSTI ID:
1474343
Report Number(s):
LLNL-JRNL-739584; 893277
Journal Information:
ACM Transactions on Embedded Computing Systems, Vol. 16, Issue 5s; ISSN 1539-9087
Publisher:
Association for Computing Machinery (ACM)Copyright Statement
Country of Publication:
United States
Language:
English
Citation Metrics:
Cited by: 2 works
Citation information provided by
Web of Science

References (5)

Pin: building customized program analysis tools with dynamic instrumentation conference January 2005
AprilTag: A robust and flexible visual fiducial system conference May 2011
Green: a framework for supporting energy-conscious programming using controlled approximation conference January 2010
Floating-Point Shadow Value Analysis conference November 2016
Analysis and characterization of inherent application resilience for approximate computing conference January 2013