Presentation is loading. Please wait.

Presentation is loading. Please wait.

Embedded Lab. Park Yeongseong.  Introduction  Problem Formulation  Approach Overview  AOI(Area Of Interest) Extraction  CallStack Pattern Mining.

Similar presentations


Presentation on theme: "Embedded Lab. Park Yeongseong.  Introduction  Problem Formulation  Approach Overview  AOI(Area Of Interest) Extraction  CallStack Pattern Mining."— Presentation transcript:

1 Embedded Lab. Park Yeongseong

2  Introduction  Problem Formulation  Approach Overview  AOI(Area Of Interest) Extraction  CallStack Pattern Mining  CallStack Pattern Clustering  Evaluation  Conclusion  Q&A

3  Modern software system ◦ Large number of components ◦ Large number of lines of code ◦ Depend on large number of system components  Recent Solution ◦ Microsoft Windows Error Reporting (WER) ◦ Performance bugs not handled (crash/hang debug)  PerfTrack : Event Tracing for Windows  Dtrace : Solaris and several other Unix-like system

4  Propose approach ◦ StackMine  StackMine ◦ To enable performance debugging in the large ◦ Support scalable system for postmortem performance debugging  StackMine include 2phases ◦ Costly-pattern mining algorithm ◦ Clustering on the mined costly pattern

5

6  Majority performance bug ◦ CPU consumption bug ◦ Wait bug

7

8

9  3 steps to reduce the investigation scope ◦ AOI extraction ◦ Costly-maximal-pattern mining ◦ CallStack pattern clustering

10  Two major issues ◦ Effectiveness ◦ Efficiency  Two effective AOI extraction techniqes ◦ Scope-based extraction ◦ Content-based extraction

11

12

13  More easily recognize the common part and the variant part of the callstack patterns.  Using performance metrics of a cluster can help produce better prioritization of results for investigation

14

15

16  Window Explorer UI  PerfTrack : collected 6000 trace stream  Randomly select 1000 trace stream  Relevant 921 trace stream( 181million callstack) ◦ 140 million : waiting stacks ◦ 41 running stacks

17  StackMine : ranked list 1215 pattern clusters.  AOI extract : 141million -> 689thousand  Maximal-callstack-pattern mining : produce 2239 costly patterns.  Final Ranked costly pattern lists : 1215  Top 400 clusters -> 93 performance signature  58.26% of the response delay time.  Average 1.6 seconds of UI response delay

18  To enable performance debugging in the large : StackMine  StackMine helps performance analysts effectively discover highly impactful performance bugs.  StackMine’s substantial benefits in performance debugging in the large

19


Download ppt "Embedded Lab. Park Yeongseong.  Introduction  Problem Formulation  Approach Overview  AOI(Area Of Interest) Extraction  CallStack Pattern Mining."

Similar presentations


Ads by Google