1 / 14

Joël Goossens Nathan Fisher

The Chilling Effect of Parallelism: Analysis and Allocation of Parallel Real-Time Jobs for Peak System-Temperature Minimization. Joël Goossens Nathan Fisher Université Libre de Bruxelles Wayne State University. Challenge: Thermal Management.

jenny
Download Presentation

Joël Goossens Nathan Fisher

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. The Chilling Effect of Parallelism: Analysis and Allocation of Parallel Real-Time Jobs for Peak System-Temperature Minimization JoëlGoossens Nathan Fisher UniversitéLibre de Bruxelles Wayne State University

  2. Challenge: Thermal Management Heatis a by-product of computation. Processors must operate within thermal thresholds: • Reliability • Safety • Cooling Costs Dynamic Voltage/Frequency Scaling (DVFS) utilized to ensure no thresholds violations.

  3. Current Research Trend: Thermal-Aware Real-Time Systems Common Objective: Minimize peak system temperature platform using DVFS cores while guaranteeing real-time constraints. Our Setting: Multicore Architectures with DVFS

  4. Our Setting: MulticoreThermal Management • Thermal Challenges: • Heat transfer between elements (e.g., Core-to-Sink). Core-to-Sink Core 3 Core 4 Sink 1 Sink 2 Core 1 Core 2

  5. Our Setting: MulticoreThermal Management • Thermal Challenges: • Heat transfer between elements (e.g., Core-to-Sink). Sink-to-Core Core 3 Core 4 Sink 1 Sink 2 Core 1 Core 2

  6. Our Setting: MulticoreThermal Management • Thermal Challenges: • Heat transfer between elements (e.g., Core-to-Sink). Sink-to-Sink Core 3 Core 4 Sink 1 Sink 2 Core 1 Core 2

  7. Our Setting: MulticoreThermal Management • Thermal Challenges: • Heat transfer between elements (e.g., Core-to-Sink). Core-to-Core Core 3 Core 4 Sink 1 Sink 2 Core 1 Core 2

  8. Our Setting: MulticoreThermal Management • Thermal Challenges: • Heat transfer between elements (e.g., Core-to-Sink). Sink-to-Environment Core 3 Core 4 Sink 1 Sink 2 Core 1 Core 2 Previous Work: Most prior work has focused on minimizing peak temperature in multicore processors for non-parallel real-time jobs. Open Question: Can parallel jobs help further minimize peak temperature?

  9. Our Setting: Parallel Real-Time Jobs “Traditional” (Sequential) Recurring Task i = (ei,di,pi): • execution requirement. • relative deadline. • minimum inter-arrival separation (“period”). • Parallel Recurring Task i= (ei,Γi,di,pi) • execution requirement. • relative deadline. • minimum inter-arrival separation (“period”). • parallel speed-up vector Γi = (γi,1, γi,2, …,γi,m) Execution on ℓ processorsfor t time units: γi,ℓx t Each parallel execution is called a “thread”

  10. Our Setting: Parallel Real-Time Jobs • Degree of Parallelism Models: • Rigid: degree determined a priori. • Moldable: chosen by scheduler at start of each job. • Malleable: may dynamically change over job execution. • Type of Parallelism Models: • Multi-Threaded: threads can execute concurrently. • Which includes Fork-Join task model. • Gang: threads must execute in unison.

  11. Motivating Example • Consider two-core processor with one task: • i= (ei,Γi,di,pi) = (1,[1,2],1,1) • Assume that processor speed is fixed at design-time. Option 1 (No Parallelism): One processor must execute at speed one. Option 2 (Degree-2 Parallelism): Each processor can execute at half-speed. Observation: Option 1 has greater peak temperature than Option 2 (even if some overhead is added for parallelism). Parallelism helps by spreading out heat generation!

  12. Open Problems Problem 1: Schedulability analysis for parallel jobs on platforms where cores run at different speeds. Problem 2: Online scheduling algorithms for thermal-aware parallel jobs. Problem 3: Core-speed assignment algorithms for DVFS-capable cores.

  13. Open Problems ? ? ? ? ? ?

  14. Thank You! Questions?

More Related