-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/31456/#review74187
-----------------------------------------------------------

Ship it!


Ship It!

- Maxim Khutornenko


On Feb. 26, 2015, 2 a.m., Bill Farner wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/31456/
> -----------------------------------------------------------
> 
> (Updated Feb. 26, 2015, 2 a.m.)
> 
> 
> Review request for Aurora, Kevin Sweeney and Maxim Khutornenko.
> 
> 
> Bugs: AURORA-1153
>     https://issues.apache.org/jira/browse/AURORA-1153
> 
> 
> Repository: aurora
> 
> 
> Description
> -------
> 
> A small, but cheap performance improvement when the preemptor is engaged.
> 
> 
> Diffs
> -----
> 
>   
> src/main/java/org/apache/aurora/scheduler/async/preemptor/PreemptorImpl.java 
> 42af883721e4e5c0ae23ff65a5fb7dc285e48faa 
> 
> Diff: https://reviews.apache.org/r/31456/diff/
> 
> 
> Testing
> -------
> 
> Before this diff:
> 
> ```
> o.a.a.b.SchedulingBenchmarks.ConstraintMismatchsSchedulingBenchmark.runBenchmark
>        avgt      100   5788175.295 ±  47410.511  ns/op
> o.a.a.b.SchedulingBenchmarks.InsufficientResourcesSchedulingBenchmark.runBenchmark
>      avgt      100   5046038.818 ±  38636.961  ns/op
> o.a.a.b.SchedulingBenchmarks.PreemptorFallbackForLargeClusterBenchmark.runBenchmark
>     avgt      100  64323073.493 ± 644053.437  ns/op
> ```
> 
> After this diff:
> ```
> o.a.a.b.SchedulingBenchmarks.ConstraintMismatchsSchedulingBenchmark.runBenchmark
>        avgt      100   5783859.480 ±  37291.662  ns/op
> o.a.a.b.SchedulingBenchmarks.InsufficientResourcesSchedulingBenchmark.runBenchmark
>      avgt      100   5259022.679 ±  30276.902  ns/op
> o.a.a.b.SchedulingBenchmarks.PreemptorFallbackForLargeClusterBenchmark.runBenchmark
>     avgt      100  58171450.071 ± 638305.308  ns/op
> ```
> 
> So, about 10% for ~free.
> 
> 
> Thanks,
> 
> Bill Farner
> 
>

Reply via email to