Name
Abstract | ||
|---|---|---|
Most programs are repetitive, where similar behavior can be seen at different execution times. Proposed algorithms au- tomatically group similar portions of a program's execution into phases, where the intervals in each phase have homo- geneous behavior and similar resource requirements. These prior techniques focus on fixed length intervals (such as a hundred million instructions) to find phase behavior. Fixed length intervals can make a program's periodic phase behav- ior difficult to find, because the fixed interval length can be out of sync with the period of the program's actual phase be- havior. In addition, a fixed interval length can only express one level of phase behavior. In this paper, we graphically show that there exists a hierar- chy of phase behavior in programs and motivate the need for variable length intervals. We describe the changes applied to SimPoint to support variable length intervals. We finally con- clude by providing an initial study into using variable length intervals to guide SimPoint. |
Year | DOI | Venue |
|---|---|---|
2005 | 10.1109/ISPASS.2005.1430568 | IEEE Electrical Insulation Magazine |
Keywords | Field | DocType |
hierarchical phase,variable length intervals,phase detection,frequency,hardware,computer science,multi threading,machine learning | Multithreading,Existential quantification,Computer science,Homogeneous,Parallel computing,Algorithm,Hierarchy,sync,Periodic graph (geometry) | Conference |
ISBN | Citations | PageRank |
0-7803-8965-4 | 39 | 1.56 |
References | Authors | |
25 | 5 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| J. Lau | 1 | 39 | 1.56 |
| E. Perelman | 2 | 89 | 4.56 |
| G. Hamerly | 3 | 89 | 4.56 |
| T. Sherwood | 4 | 62 | 3.49 |
| Brad Calder | 5 | 4145 | 251.59 |