Skip to content

adjust readpool size by the number of cpu core#5654

Merged
zhangjinpeng87 merged 2 commits intotikv:masterfrom
zhangjinpeng87:tune-default-poo-size
Oct 15, 2019
Merged

adjust readpool size by the number of cpu core#5654
zhangjinpeng87 merged 2 commits intotikv:masterfrom
zhangjinpeng87:tune-default-poo-size

Conversation

@zhangjinpeng87
Copy link
Member

@zhangjinpeng87 zhangjinpeng87 commented Oct 15, 2019

Signed-off-by: zhangjinpeng1987 zhangjinpeng@pingcap.com

What have you changed?

Adjust readpool size by the number of cpu core, make it works better under sysbench and other benchmark cases when using default configuration value.

What is the type of the changes?

Pick one of the following and delete the others:

  • Improvement (a change which is an improvement to an existing feature)

How is the PR tested?

  • Manual test (add detailed scripts or steps below)

  • benchmark

Does this PR affect documentation (docs) or should it be mentioned in the release notes?

  • If there is a document change, please file a PR in (docs) and add the PR number here.
  • If this PR should be mentioned in the release note, please update the release notes.

Does this PR affect tidb-ansible?

No

Signed-off-by: zhangjinpeng1987 <zhangjinpeng@pingcap.com>
@zhangjinpeng87 zhangjinpeng87 added the component/performance Component: Performance label Oct 15, 2019
@zhangjinpeng87
Copy link
Member Author

/bench

Signed-off-by: zhangjinpeng1987 <zhangjinpeng@pingcap.com>
@zhangjinpeng87
Copy link
Member Author

/bench

Copy link
Member

@breezewish breezewish left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Documentations need to be updated accordingly.

@breezewish
Copy link
Member

/run-integration-cop-push-down-test tidb-test=pr/899

@sre-bot
Copy link
Contributor

sre-bot commented Oct 15, 2019

@@                               Benchmark Diff                               @@
================================================================================
tidb: 1fe25e071cb02bafce8eab288838d412ef358556
--- tikv: 6f4bc8cc29f6bdcaac8f4c2b975ef994abfb3fac
+++ tikv: 41b3e9c8b2fbb2745e49ef4275509a25955382cb
pd: 82bb13a942e32f2fc6311fea66d460586a8924a0
================================================================================
test-1: < oltp_point_select >
    * QPS : 138877.25 ± 0.1438% (std=155.53) delta: 43.73% (p=0.000)
    * AvgMs : 1.84 ± 0.0000% (std=0.00) delta: -30.57%
    * PercentileMs99 : 6.21 ± 0.0000% (std=0.00) delta: -1.74%
            
test-2: < oltp_read_write >
    * QPS : 41579.39 ± 0.4690% (std=123.24) delta: 0.11% (p=0.358)
    * AvgMs : 123.61 ± 0.4692% (std=0.36) delta: -0.12%
    * PercentileMs99 : 248.83 ± 0.0000% (std=0.00) delta: 1.82%
            
test-3: < oltp_insert >
    * QPS : 23809.54 ± 0.3289% (std=46.00) delta: -0.16% (p=0.885)
    * AvgMs : 10.75 ± 0.3164% (std=0.02) delta: 0.17%
    * PercentileMs99 : 21.89 ± 0.0000% (std=0.00) delta: 0.00%
            
test-4: < oltp_update_index >
    * QPS : 19242.12 ± 0.1928% (std=27.12) delta: 0.97% (p=0.282)
    * AvgMs : 13.28 ± 0.3915% (std=0.04) delta: 0.05%
    * PercentileMs99 : 30.81 ± 0.0000% (std=0.00) delta: -0.00%
            
test-5: < oltp_update_non_index >
    * QPS : 32370.31 ± 0.2018% (std=37.23) delta: 1.03% (p=0.002)
    * AvgMs : 7.90 ± 0.2024% (std=0.01) delta: -1.03%
    * PercentileMs99 : 20.52 ± 1.0820% (std=0.18) delta: -1.07%
            

https://perf.pingcap.com

Copy link
Member

@tabokie tabokie left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

component/performance Component: Performance

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants