Skip to content
This repository has been archived by the owner on Jan 13, 2023. It is now read-only.

Start using wrk instead of ab for all benchmarks #30

Closed
2 tasks done
geerlingguy opened this issue Feb 26, 2015 · 2 comments
Closed
2 tasks done

Start using wrk instead of ab for all benchmarks #30

geerlingguy opened this issue Feb 26, 2015 · 2 comments

Comments

@geerlingguy
Copy link
Owner

From some informal testing, it looks like wrk might be a little more lightweight and accurate than ab for local-network tests. Plus the summary is much more concise, which I like :)

For example:

$ wrk -t12 -c200 -d30 http://pidramble.com/about
Running 30s test @ http://pidramble.com/about
  12 threads and 200 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency     4.37s     6.33s   16.36s    79.36%
    Req/Sec   417.69    650.51     5.44k    92.71%
  78676 requests in 30.01s, 516.52MB read
  Socket errors: connect 0, read 0, write 0, timeout 652
Requests/sec:   2622.02
Transfer/sec:     17.21MB

I need to:

  • Update README.md docs for benchmarking.
  • Re-run authenticated/cached/uncached benchmarks.
@geerlingguy
Copy link
Owner Author

That's more like it... getting much nicer looking numbers now (on local Vagrant VMs):

Cached:

$ wrk -t4 -c100 -d30 http://pidramble.com/about
Running 30s test @ http://pidramble.com/about
  4 threads and 100 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency    81.71ms  167.94ms 672.44ms   90.24%
    Req/Sec     0.88k   424.31     3.07k    80.13%
  102027 requests in 30.00s, 669.82MB read
Requests/sec:   3400.89
Transfer/sec:     22.33MB

Uncached:

$ wrk -t4 -c4 -d10 http://pidramble.com/about?nocache=true
Running 10s test @ http://pidramble.com/about?nocache=true
  4 threads and 4 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency   236.36ms   84.98ms 515.88ms   56.00%
    Req/Sec     3.96      1.54     7.00     58.67%
  177 requests in 10.02s, 1.16MB read
Requests/sec:     17.67
Transfer/sec:    118.63KB

Authenticated:

Note: It takes a little more configuration to use wrk with cookies...

@geerlingguy
Copy link
Owner Author

And on the actual cluster itself:

Cached:

$ wrk -t4 -c100 -d30 http://pidramble.com/about
Running 30s test @ http://pidramble.com/about
  4 threads and 100 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency    65.41ms   28.85ms 147.46ms   73.17%
    Req/Sec   387.76     47.92   457.00     66.77%
  46683 requests in 30.00s, 304.74MB read
Requests/sec:   1555.94
Transfer/sec:     10.16MB

Uncached:

$ wrk -t1 -c100 -d10 http://pidramble.com/about?nocache=true
Running 10s test @ http://pidramble.com/about?nocache=true
  1 threads and 100 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency     5.59s     2.16s    8.97s    71.67%
    Req/Sec    11.00      0.00    11.00    100.00%
  108 requests in 10.02s, 720.62KB read
  Socket errors: connect 0, read 0, write 0, timeout 320
Requests/sec:     10.77
Transfer/sec:     71.88KB

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant