mesos-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Benjamin Mahler <bmah...@apache.org>
Subject Re: Review Request 70430: Avoided some recalculation in the random sorter.
Date Thu, 25 Apr 2019 19:13:07 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/70430/#review214891
-----------------------------------------------------------


Fix it, then Ship it!




Maybe we can re-run the benchmark and include the results in the commit since we didn't have
the dirty bit getting cleared?


src/master/allocator/sorter/random/sorter.cpp
Line 629 (original), 643 (patched)
<https://reviews.apache.org/r/70430/#comment301181>

    set dirty bit to false?


- Benjamin Mahler


On April 9, 2019, 12:01 a.m., Meng Zhu wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/70430/
> -----------------------------------------------------------
> 
> (Updated April 9, 2019, 12:01 a.m.)
> 
> 
> Review request for mesos and Benjamin Mahler.
> 
> 
> Bugs: MESOS-9733
>     https://issues.apache.org/jira/browse/MESOS-9733
> 
> 
> Repository: mesos
> 
> 
> Description
> -------
> 
> This patch keeps the sorting related information in the memory
> and accompanies a dirty bit with it. This helps to avoid
> unnecessary recalculation of this info in `sort()`.
> 
> 
> Diffs
> -----
> 
>   src/master/allocator/sorter/random/sorter.hpp 125ce84761e4c930370912151700ddda35d7b6c1

>   src/master/allocator/sorter/random/sorter.cpp bbe130dbf3b158ea14f9572bc5d14200fcd85127

> 
> 
> Diff: https://reviews.apache.org/r/70430/diff/3/
> 
> 
> Testing
> -------
> 
> make check
> 
> Benchmarking:
> Optimized build with QuotaParam/BENCHMARK_HierarchicalAllocator_WithQuotaParam.LargeAndSmallQuota/2
> 
> tl;dr: Performance is now on par with before the change in r/70429
> 
> ## Before r/70429:
> Added 3000 agents in 85.844373ms
> Added 3000 frameworks in 19.713969615secs
> Benchmark setup: 3000 agents, 3000 roles, 3000 frameworks
> Made 3500 allocations in 13.690538305secs
> Made 0 allocation in 9.76855825secs
> 
> ## After r/70429 and this:
> Added 3000 agents in 105.284907ms
> Added 3000 frameworks in 19.253697264secs
> Benchmark setup: 3000 agents, 3000 roles, 3000 frameworks
> Made 3500 allocations in 13.488219803secs
> Made 0 allocation in 9.846584089secs
> 
> 
> Thanks,
> 
> Meng Zhu
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message