Understanding Energy Behaviors of Thread Management Constructs

Transcript

1. Understanding Energy Behaviors of Thread Management Constructs Gustavo Pinto1 Fernando

   Castor1 David Liu2 {ghlp, castor}@cin.ufpe.br1 davidL@binghamton.edu2

2. 2 • First, the proliferation of multicore CPUs • Second,

   the prevalence of multi-threaded programs Motivation

3. 3 • First, more cores more power consumed • Second,

   little is known about energy behaviors of multi-threaded program on the application and programming language level The Problem

4. 4 1. programming abstractions of thread management on energy efficiency

   2. programmer choices of thread management on energy efficiency This Talk for Java multi-threaded programs

5. 5 • Explicit threading (the Thread-style): Using the java.lang.Thread class

   • Thread pooling (the Executor-style): Using the java.util.concurrent.Executor framework • Working Stealing (the ForkJoin-style): Using the java.util.concurrent.ForkJoin framework Thread management constructs

6. 6 • Embarrassingly parallel: spectralnorm, sunflow, n-queens • Leaning parallel:

   xalan, knucleotide, tomcat • Leaning serial: mandelbrot, largestImage • Embarrassingly serial: h2 Benchmarks

7. 7 • Embarrassingly parallel: spectralnorm, sunflow, n-queens • Leaning parallel:

   xalan, knucleotide, tomcat • Leaning serial: mandelbrot, largestImage • Embarrassingly serial: h2 Benchmarks Micro-benchmarks DaCapo benchmarks

8. 8 Experimental Environment A 2×16-core AMD CPUs, running Debian Linux,

   64GB of memory, JDK version 1.7.0 11, build 21, “ondemand” governor

9. 9 Experimental Environment A 2×16-core AMD CPUs, running Debian Linux,

   64GB of memory, JDK version 1.7.0 11, build 21, “ondemand” governor

10. 10 Experimental Environment A 2×16-core AMD CPUs, running Debian Linux,

    64GB of memory, JDK version 1.7.0 11, build 21.

11. 11 Experimental Environment A 2×16-core AMD CPUs, running Debian Linux,

    64GB of DDR3 1600 memory, and JDK version 1.7.0 11, build 21.

12. Energy Consumption When Varying the Number of Threads 12

13. 13 The Λ Curve

14. 14 The Λ Curve

15. 15 The Λ Curve

16. 16 More cores idle CPU frequency at a lower level

    The Λ Curve

17. 17 More cores idle CPU frequency at a lower level

    More threads used, program completes sooner The greater the ratio between speedup and power, the steeper the \ The Λ Curve

18. Which programming style should I use? 18 ? ?

19. Overpopulating Cores with Threads 19

20. Overpopulating Cores with Threads 20

21. Faster ≠ Greener 21

22. Copying vs Sharing 22

23. Copying vs Sharing 23 Copying

24. Copying vs Sharing 24 Copying Sharing

25. Copying vs Sharing 25 Copying Sharing ±15% of energy savings!

26. 26 Copy-Fork

27. 27 After Before Copy-Fork

28. 28 After Before Copy-Fork Copy/Fork/Copy/Fork/…

29. 29 After Before Copy-Fork Copy/Fork/Copy/Fork/…

30. 30 After Before Copy-Fork Copy/Fork/Copy/Fork/… Copy/../Copy/Fork/…/Fork

31. 31 After Before Copy-Fork ±10% of energy savings! Copy/Fork/Copy/Fork/… Copy/../Copy/Fork/…/Fork

32. Data Size 32

33. Data Size 33 For most of the cases, energy consumption

    is linear to data size!

34. 34

35. 35

36. 36

37. 37

38. Understanding Energy Behaviors of Thread Management Constructs Gustavo Pinto1 Fernando

    Castor1 David Liu2 {ghlp, castor}@cin.ufpe.br1 davidL@binghamton.edu2