12
Progress Report 2012/12/20

Progress Report

  • Upload
    trula

  • View
    34

  • Download
    0

Embed Size (px)

DESCRIPTION

Progress Report. 2012/12/20. Computation Offloading. Mobile devices have limited energy and computing resources. Offloading some workloads to remote servers leads to: Power-saving. Performance improving. Shorter execution time. Better results. Our Idea. - PowerPoint PPT Presentation

Citation preview

Page 1: Progress Report

Progress Report2012/12/20

Page 2: Progress Report

Computation OffloadingMobile devices have limited

energy and computing resources.Offloading some workloads to

remote servers leads to:◦Power-saving.◦Performance improving.

Shorter execution time. Better results.

Page 3: Progress Report

Our IdeaInstead of studying offloading

policies, we aim at the effects caused by offloading.

After offloading, a computation-intensive process becomes I/O-intensive.◦Does this phenomenon affects

scheduler?◦Does this phenomenon affects DVFS?◦Does this phenomenon affects

cache/memory?◦…etc.

Page 4: Progress Report

ProblemService thread

◦Computation-intensive => I/O-intensive

◦Task in waiting state will not be scheduled.

Page 5: Progress Report

Current Flow

Offloading Framework

Scheduler DVFS

DPMLoad

change

Page 6: Progress Report

New strategyThe original idea of virtual core

is:

N cores

C-task

Remote cores

Computation Offloading

N+1 cores

Page 7: Progress Report

New strategy(Cont.)反其道而行Close a core after offloading

N cores

C-task

Remote cores

Computation Offloading

N-1 cores

Page 8: Progress Report

ReasonAfter offloading, the rest of the

tasks will be scheduled to N cores.◦Should have better performance.◦Does not guarantee energy saving.

All the cores are still working!

If we close a core after offloading◦Imagine that the computation-

intensive task is non-preemptive, but consume zero power.

◦Energy saving with little effect to (other tasks) performance.

Page 9: Progress Report

New Flow

Offloading Framework

Scheduler DVFS

DPM

Page 10: Progress Report

Comparison

Performance of other tasks

Power Consumption

No offloading -- Highest

Virtual core (N+1) Better than No offloading

Low

Proposed strategy (N-1)

Similar to No offloading

Lowest

Page 11: Progress Report

HoweverThis is a theoretically strategy.Need to design some

experiments to verify the strategy.

Page 12: Progress Report

Possible TopicsClose more cores

◦-1?-2?-n/2?-(n-1)?Close which core(s)

◦The one executing the offloaded task?

◦Cache relatedTheoretical problem

◦Math model