1 / 19

MARS: Adaptive Remote Execution Scheduler for Multithreaded Mobile Devices

MARS: Adaptive Remote Execution Scheduler for Multithreaded Mobile Devices. Asaf Cidon *, Tomer M. London*, Sachin Katti , Christos Kozyrakis , Mendel Rosenblum. Stanford University. *Equal contributors. New Class of Mobile Applications. Computer Vision. Motion Sensing.

robbin
Télécharger la présentation

MARS: Adaptive Remote Execution Scheduler for Multithreaded Mobile Devices

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. MARS: Adaptive Remote Execution Scheduler for Multithreaded Mobile Devices AsafCidon*, Tomer M. London*, SachinKatti, Christos Kozyrakis, Mendel Rosenblum Stanford University *Equal contributors

  2. New Class of Mobile Applications Computer Vision Motion Sensing Augmented Reality

  3. Mobile Client Trends • Mobile CPU performance increasing • Hitting ‘energy wall’ • Can we improve performance and reduce energy consumption? • Opportunity: network bandwidth increase utilize the cloud Maximum Bandwidth (Mb/s)

  4. Static Client-Server PartitioningDoesn’t Work • Dynamic resources: • Network bandwidth and latency • Available CPU, memory • Same code, different platforms: • Smartphones (single-core, multi-core) • Tablets

  5. MARS: Adaptive Remote Execution • Opportunistically offload computations to remote server • Enhance computational capabilities • Decrease energy consumption • Make dynamic decisions • Adapt to network and CPU variability Mobile Device Data Center

  6. Agenda • Design of MARS • Simulator Results and Analysis • Conclusions

  7. Existing Remote Execution Systems The Unit of Remote Execution Cloudlets [Satyanarayanan et al., ‘09] CloneCloud[Kirsch et al., ‘11] VM MAUI [Cuervo et al. ‘10] MARS “Cloud-on-Chip” Odessa [Ra et al. ‘11] RPC Chroma [Balan et al. ‘03] Target of Performance Optimization Single-thread application Multi-threaded application System

  8. MARS “Cloud-on-Chip”: System Scheduling Previous Systems: Application Partitioning RPC Queue Local Execution Remote Execution Local Cores RPC 1 Process 1 RPC 1 Process 1 RPC 1 Process 2 RPC 2 Process 1 RPC 2 Process 1 RPC 1 Process 3 RPC 3 Process 1 RPC 4 Process 1 Remote Cores RPC 2 Process 3 RPC 5 Process 1

  9. Greedy Algorithm EOR ≥ ? Higher POR: better performance gain from offloading Higher EOR: better energy saving from offloading EOR < ?

  10. Controller Algorithm Remote Server Available RPC 3 (POR 2.5) Check EOR Threshold RPC 5 (POR 1.9) Priority Queue, sorted by Performance Offload Rank (POR) RPC 6 (POR 1.8) RPC 6 (POR 1.8) EOR Local Both Remote RPC 4 (POR 1.3) RPC 2 (POR 0.4) G (Greediness) trades-off utilization and energy efficiency Local Core Available

  11. Agenda • Design of MARS • Simulator Results and Analysis • Conclusions

  12. Remote Execution Applications Augmented Reality Face Recognition Pic Pic Pic Pic Pic Pic Barcode Detection Barcode Detection Barcode Detection Rendering Recognition Rendering Recognition Rendering Recognition

  13. Simulator Methodology • Trace-driven simulation • Clients: • Nokia N900 (single core) • NVIDIA Tegra 250 (multicore) • Server: • Amazon EC2 Opteron 2007 • Networks: • Outdoors Wi-Fi • Indoors Wi-Fi • 3G

  14. MARS vs. Static Policies

  15. Nokia N900 Power Consumption • WiFi: Performance and energy are highly correlated • 3G: trade-off performance and energy

  16. Same Application, Different Networks

  17. Remote Execution with Multicore

  18. Agenda • Design of MARS • Simulator Results and Analysis • Conclusions

  19. Conclusions • Can’t always be greedy • Performance and energy trade-off • MARS is optimized for multiple parallel applications and cores • MARS “Cloud-on-Chip”: validation of system-level remote execution scheduling • 57% performance increase, 33% energy savings

More Related