By Topic

Performance Characterization of SPEC CPU2006 Integer Benchmarks on x86-64 Architecture

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

4 Author(s)
Dong Ye ; Dept. of ECE, Northeastern Univ., Boston, MA ; Ray, J. ; Harle, C. ; Kaeli, D.

As x86-64 processors become the CPU of choice for the personal computer market, it becomes increasingly important to understand the performance we can expect by migrating applications from a 32-bit environment to a 64-bit environment. For applications that can effectively exploit a larger memory address space (e.g., commercial databases and digital content authoring tools), it is not surprising that x86-64 can provide a performance boost. However, for less-demanding desktop applications that can fit in a 32-bit address space, we would like to know if we can expect any performance benefits by moving to this platform. In this paper, we report on a range of performance characteristics for programs compiled for both 32 bits and 64 bits and run directly (32-bit binaries are run in compatibility mode; 64-bit binaries are run in 64-bit mode) on a single x86-64 based system. In this study we utilize the integer benchmarks from the newly released SPEC CPU2006 suite. We have observed that for the SPEC CPU2006 integer benchmarks, 64-bit mode offers a sizable performance advantage over 32-bit mode (7% on average). However, the advantages vary from benchmark to benchmark, and for a handful of programs, 64-bit mode is significantly slower than 32-bit mode (in this subset of benchmarks, performance is reduced by more than 16% when running in 64-bit mode.) We further analyze 5 benchmarks that exhibit significant differences in performance between these two modes. For this set of CPU2006 integer programs, we present a range of performance characteristics that illustrate the impact of moving to a 64-bit environment. Our results and analysis can be used by performance engineers and developers to better understand how to exploit the capabilities of the x86-64 architecture

Published in:

Workload Characterization, 2006 IEEE International Symposium on

Date of Conference:

25-27 Oct. 2006