Skip to content

Instantly share code, notes, and snippets.

@ruizander
Created Nov 17, 2015
Embed
What would you like to do?
benchmark enabling different hyper version (keep_alive enabled by default on 0.6.14)
my original Cargo.toml:
[package]
name = "nickel-learn"
version = "0.1.0"
authors = ["Ruizander <ruizander@gmail.com>"]
[dependencies]
nickel = "*"
===============================================================================================
wrk -t4 -c100 -d30S --timeout 2000 http://127.0.0.1:6767 [13/2628]
Running 30s test @ http://127.0.0.1:6767
4 threads and 100 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 25.44ms 43.44ms 373.65ms 93.51%
Req/Sec 1.16k 592.18 2.64k 65.25%
32850 requests in 30.08s, 5.20MB read
Requests/sec: 1092.23
Transfer/sec: 177.06KB
===============================================================================================
changing my Cargo.toml to:
[package]
name = "nickel-learn"
version = "0.1.0"
authors = ["Ruizander <ruizander@gmail.com>"]
[dependencies]
hyper = "=0.6.14"
nickel = "*"
=================================================================================
and repeating the same workload:
wrk -t4 -c100 -d30S --timeout 2000 http://127.0.0.1:6767
Running 30s test @ http://127.0.0.1:6767
4 threads and 100 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 128.79us 437.22us 93.46ms 99.41%
Req/Sec 14.45k 1.88k 16.12k 89.37%
432722 requests in 30.10s, 60.66MB read
Requests/sec: 14376.08
Transfer/sec: 2.02MB
...that would be ~1300% more requests?!?
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment