A hardware mechanism for dynamic extraction and relayout of program hot spots

Merten, Matthew C.; Trick, Andrew R.; Nyström, Erik; Barnes, R. D.; Hmu, Wen-mei W.

doi:10.1145/339647.339655

Cited by 29 publications

(28 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(1,1), and 69% for CBS (16,3) for 1 to 2% overhead [4]. We show better base results here with an average accuracy of 52% for CBS (1,1) and 76% for CBS (16,3).…”

Section: Accuracymentioning

confidence: 52%

“…To compare the effect of the sampling configuration on call graph correction, we use two sampling configurations: CBS (1,1) and CBS (16,3), 16 samples with a stride of 3. The default sampling configuration in Jikes RVM is CBS (1,1), which is equivalent to the default timer-based sampling in Figure 1(a).…”

Section: Resultsmentioning

confidence: 99%

“…The default sampling configuration in Jikes RVM is CBS (1,1), which is equivalent to the default timer-based sampling in Figure 1(a). Arnold and Grove recommend CBS (16,3), which takes more samples to increase accuracy while keeping average overhead down to 1-2%.…”

Section: Resultsmentioning

confidence: 99%

“…FDOM and intraprocedural correction are most effective when the base graph is less accurate as in CBS(1,1) because they improve relative frequencies within a method. Interprocedural correction is relatively more effective using a more accurate base graph such as CBS (16,3). This result is intuitive; a global scheme for improving accuracy works best when its constituent components are accurate.…”

Section: Accuracymentioning

confidence: 94%

“…Most dynamic optimizers collect accurate control-flow profiles such as basic block and edge profiles to make better optimization decisions [1,3,12,16,17]. We show how to combine these constraints to further improve the accuracy of the DCG.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Correcting the Dynamic Call Graph Using Control-Flow Constraints

Lee

Resnick

Bond

et al.

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. To reason about programs, dynamic optimizers and analysis tools use sampling to collect a dynamic call graph (DCG). However, sampling has not achieved high accuracy with low runtime overhead. As object-oriented programmers compose increasingly complex programs, inaccurate call graphs will inhibit analysis and optimizations. This paper demonstrates how to use static and dynamic control flow graph (CFG) constraints to improve the accuracy of the DCG. We introduce the frequency dominator (FDOM), a novel CFG relation that extends the dominator relation to expose static relative execution frequencies of basic blocks. We combine conservation of flow and dynamic CFG basic block profiles to further improve the accuracy of the DCG. Together these approaches add minimal overhead (1%) and achieve 85% accuracy compared to a perfect call graph for SPEC JVM98 and DaCapo benchmarks. Compared to sampling alone, accuracy improves by 12 to 36%. These results demonstrate that static and dynamic control-flow information offer accurate information for efficiently improving the DCG.

show abstract

“…(1,1), and 69% for CBS (16,3) for 1 to 2% overhead [4]. We show better base results here with an average accuracy of 52% for CBS (1,1) and 76% for CBS (16,3).…”

Section: Accuracymentioning

confidence: 52%