By Topic

Victim management in a cache hierarchy

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
$33 $31
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)
Franaszek, P.A. ; IBM Research Division, Thomas J. Watson Research Center, P.O. Box 218, Yorktown Heights, New York 10598, USA ; Lastras-Montaño, L.A. ; Kunkel, S.R. ; Sawdey, A.C.

We investigate directions for exploiting what might be termed pattern locality in a cache hierarchy, based on recording cache discards or victims. An advantage of storing discard decisions is the reduced duplication of pertinent information, as well as the maintenance of information on the current location of discarded lines. Typical caches are designed to exploit combinations of temporal and spatial locality. Temporal locality, the likelihood that recently referenced data will be referenced again, is exploited by LRU-like algorithms. Spatial locality is the property that causes larger cache lines to yield improved miss ratios. Here we consider the exploitation of pattern locality—the property that lines accessed in temporal proximity tend to be re-referenced together. We describe some new cache structures including pattern-recording features, along with their miss ratio and transfer traffic performance as determined via simulations on traces drawn from several benchmark applications. We show that pattern locality information, based on discard statistics, can be useful in enhancing the quality of prefetch decisions.

Note: The Institute of Electrical and Electronics Engineers, Incorporated is distributing this Article with permission of the International Business Machines Corporation (IBM) who is the exclusive owner. The recipient of this Article may not assign, sublicense, lease, rent or otherwise transfer, reproduce, prepare derivative works, publicly display or perform, or distribute the Article.  

Published in:

IBM Journal of Research and Development  (Volume:50 ,  Issue: 4.5 )