Presentation is loading. Please wait.

Presentation is loading. Please wait.

Progress Report 2015/08/31. Previously Energy Aware Scheduling Kernel Summit 2014 ◦ rt-app: creates workload patterns based on workload description.

Similar presentations


Presentation on theme: "Progress Report 2015/08/31. Previously Energy Aware Scheduling Kernel Summit 2014 ◦ rt-app: creates workload patterns based on workload description."— Presentation transcript:

1 Progress Report 2015/08/31

2 Previously Energy Aware Scheduling workshop @ Kernel Summit 2014 ◦ rt-app: creates workload patterns based on workload description files (current JSON).

3 rt-app A test application that starts multiple periodic threads in order to simulate a real-time periodic load. Not only the sleep and run pattern can be emulated, but also the dependency between tasks. ◦ accessing same critical resources, creating sequential wake up or syncing the wake up of threads. The use case is described in a json like file https://git.linaro.org/power/rt-app.git/

4 Workloads Main objects ◦ Global objects  Duration, Calibration, default_policy, io_device, … ◦ Tasks/Threads  Instance, priority, cpus, phases, …  run, sleep, mem, iorun, timer, lock, wait, sync, …

5 Run event Emulate the execution of a load. ◦ The duration is defined in usec, but the run event will effectively run a number of time a loop that waste cpu cycles. ◦ The duration is transformed in a number of loop by the ns per loop value defined in calibration.  Default: CPU[0] ◦ The actual duration will vary with the frequency or the compute capacity of the CPU.

6 Outputs Log some metrics for each phase that is executed by a thread ◦ perf: number of loop that has been executed by run events ◦ run: time spent by the thread to execute the run events ◦ period: duration to execute the complete phase ◦ start/end : absolute start and end time of a phase. ◦ rel_st: start time of a phase relatively to the begin of the use case.

7 Example Output – web-BrowserMain-0.log # Policy : SCHED_OTHER priority : 0 #idx perf run period start end rel_st 0 162790 8242 16314 342761303269 342761319583 558 0 69767 3528 3531 342761319590 342761323121 16879 0 69767 3181 3184 342761323123 342761326307 20412 0 69767 3025 3028 342761326308 342761329336 23598 0 69767 3023 3026 342761329337 342761332363 26627 0 69767 3020 3023 342761332365 342761335387 29654 0 69767 3027 3030 342761335388 342761338419 32678 0 69767 3024 3028 342761338420 342761341448 35709 0 69767 3024 3026 342761341449 342761344475 38738 0 69767 3020 3023 342761344477 342761347500 41766 0 69767 3032 3035 342761347501 342761350536 44790 0 69767 3024 3027 342761350537 342761353564 47826 0 69767 3023 3026 342761353565 342761356591 50854 0 69767 3020 3023 342761356592 342761359615 53882 0 69767 3027 3030 342761359616 342761362646 56906 …

8 Examples [power/rt-app.git] / doc / examples / ◦ browser-long.json ◦ browser-short.json ◦ mp3-long.json ◦ mp3-short.json ◦ spreading-tasks.json ◦ video-long.json ◦ video-short.json

9 Next cpufreq_governor_efficiency ◦ Measure the efficiency of cpufreq governors using rt-app Cross-compile rt-app and try it on Juno board


Download ppt "Progress Report 2015/08/31. Previously Energy Aware Scheduling Kernel Summit 2014 ◦ rt-app: creates workload patterns based on workload description."

Similar presentations


Ads by Google