VRQ 0.89b is released with
1. Fix low workload regression comparing to previous vrq release by not trying to pick tasks from other run queue when cpu is scaling.
2. Follow cpu affinity order when pick tasks from other run queue.
3. Fix long existed wrong run queue scaling value when run on performance governor or exit from dynamic governor.
With all above changes, this release show better sanity performance than any other previous vrq releases. Next release will be focus on task policy fairness.
code are available at
https://bitbucket.org/alfredchen/linux-gc/commits/branch/linux-4.8.y-test
and also
https://github.com/cchalpha/linux-gc/commits/linux-4.8.y-test
all-in-one patch also available.
Enjoy it, :)
Hi!
ReplyDeletei have tested this for ~ 15 hrs now, nothing unusual, did not hang, is responsive, works fine except that C states issue.
As I mainly use laptop and power consumption is a great deal there, I'll not be testing it anymore.
If You need smth specific to test, please write.
Br, Eduardo
Thanks for testing. As there is cpu freq governor related changes, I'd like to check whether D3 playing is still fine on ondemand and performance governors.
DeleteFor C states problem, it will be taken care next week when the on going polciy fairness tests are done.
BR Alfred
@Alfred,
DeleteI tested this version, results are there (as ususal):
https://docs.google.com/spreadsheets/d/1EayezAsGlJdXjZbS3b9m7YtvtRF-DJ3xrT3hYCvfymQ/edit?usp=sharing
About D3 it's good news, no stutters, performs very well.
Br, Eduardo
Hello, and thanks for your work.
ReplyDeleteI've ran some benchmarks of VRQ0.89b out of curiosity.
Here are the results
http://openbenchmarking.org/result/1611230-LO-CFSVSMUQS36
I know it's still in early development but frequency signaling is broken with intel-pstate, giving low throughput. It's fine with acpi-cpufreq ondemand.
I'll keep an eye on future releases.
Pedro
Thanks for these overall system benchmarks, I'd like to have such benchmarks but it's too time consume for myselft.
DeleteAnd 0.89c will be released in a day or two which fixed major unexpected design intention when creating new tasks. The commit which introduced C-state issue has been found, most likely it's cpufreq_trigger related, and I am planning to fix it in 0.89d, hopefully it also help with intel-pstate or maybe the new schedutil governer.
BR Alfred
Good shaare
ReplyDelete