DocumentCode
3376109
Title
Characterizing the impact of predicated execution on branch prediction
Author
Mahlke, Scott A. ; Hank, Richard E. ; Bringmann, Roger A. ; Gyllenhaal, John C. ; Gallagher, David M. ; Hwu, Wen-Mei W.
Author_Institution
Center for Reliable & High Performance Comput., Illinois Univ., Urbana, IL, USA
fYear
1994
fDate
30 Nov.-2 Dec. 1994
Firstpage
217
Lastpage
227
Abstract
Branch instructions are recognized as a major impediment to exploiting instruction level parallelism. Even with sophisticated branch prediction techniques, many frequently executed branches remain difficult to predict. An architecture supporting predicated execution may allow the compiler to remove many of these hard-to-predict branches, reducing the number of branch mispredictions and thereby improving performance. We present an in-depth analysis of the characteristics of those branches which are frequently mispredicted and examine the effectiveness of an advanced compiler to eliminate these branches. Over the benchmarks studied, an average of 27% of the dynamic branches and 56% of the dynamic branch mispredictions are eliminated with predicated execution support.
Keywords
instruction sets; program compilers; branch instructions; branch prediction; compiler; hard-to-predict branches; instruction level parallelism; predicated execution; Character recognition; Computer aided instruction; Computer architecture; Concurrent computing; Degradation; Hardware; Impedance; Parallel processing; Permission; Processor scheduling;
fLanguage
English
Publisher
ieee
Conference_Titel
Microarchitecture, 1994. MICRO-27. Proceedings of the 27th Annual International Symposium on
ISSN
1072-4451
Print_ISBN
0-89791-707-3
Type
conf
DOI
10.1109/MICRO.1994.717461
Filename
717461
Link To Document