gistfile1.txt

## Server

EC2 c4.8xlarge
Amazon Linux

## Software

```
sudo yum install -y make gcc git curl tar bzip2 patch nginx openssl-devel strace
git clone https://github.com/tagomoris/xbuild.git
./xbuild/perl-install 5.22.0-RC2 ~/local/perl-5.22 -j30
export PATH=/home/ec2-user/local/perl-5.22/bin:$PATH
cpanm -n Gazelle Starlet HTTP::Parser::XS Starman Cache::Memcached::Fast JSON::XS
```

### wrk

```
git clone https://github.com/wg/wrk.git
cd wrk
make
```

## tune kernel

```
sudo sysctl -w net.core.netdev_max_backlog=8192
sudo sysctl -w net.core.somaxconn=32768
sudo sysctl -w net.ipv4.tcp_tw_recycle=1
sudo sysctl -w net.ipv4.tcp_tw_reuse=1
sudo sysctl -w net.ipv4.tcp_fin_timeout=2
```

# benchmark

## Hello world

### Gazelle

```
[ec2-user@ip- ~]$ start_server --backlog=16384 --path=/dev/shm/app.sock  -- plackup -s Gazelle    --max-workers 8 --max-reqs-per-child 500000 -E prod    -e 'sub{[200,[],["hello world\n"]]}'

[ec2-user@ip- wrk]$ ./wrk -t 4 -c 500 -d 30 http://localhost/
Running 30s test @ http://localhost/
  4 threads and 500 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency     2.89ms   18.07ms 393.39ms   99.24%
    Req/Sec    84.15k    11.76k  102.21k    89.42%
  10045243 requests in 30.01s, 1.41GB read
Requests/sec: 334752.37
Transfer/sec:     48.19MB
```

### Starlet

```
[ec2-user@ip- ~]$ start_server --backlog=16384 --path=/dev/shm/app.sock  -- plackup -s Starlet    --max-workers 8 --max-reqs-per-child 500000 -E prod    -e 'sub{[200,[],["hello world\n"]]}'

[ec2-user@ip- wrk]$ ./wrk -t 4 -c 500 -d 30 http://localhost/
Running 30s test @ http://localhost/
  4 threads and 500 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency     6.00ms    9.34ms 230.74ms   99.46%
    Req/Sec    23.13k     2.28k   27.40k    81.00%
  2761130 requests in 30.00s, 397.48MB read
Requests/sec:  92026.70
Transfer/sec:     13.25MB
```

### starman

```
[ec2-user@ip- ~]$ starman --listen /dev/shm/app.sock --backlog=16384 --workers 8 --max-requests 500000 -E prod --preload-app --disable-keepalive -e 'sub{[200,[],["hello world\n"]]}'

[ec2-user@ip- wrk]$ ./wrk -t 4 -c 500 -d 30 http://localhost/
Running 30s test @ http://localhost/
  4 threads and 500 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency    10.81ms    7.90ms 193.00ms   91.54%
    Req/Sec    12.29k     2.65k   17.58k    65.83%
  1467607 requests in 30.02s, 211.27MB read
Requests/sec:  48887.84
Transfer/sec:      7.04MB
```

## counter

### psgi

```
use Plack::Builder;
use Plack::Request;
use Cache::Memcached::Fast;
use JSON::XS;

my $_JSON = JSON::XS->new->utf8;
my $mem = Cache::Memcached::Fast->new({servers=>[qw/127.0.0.1:11211/]});
$mem->set('counter:1',0);
$mem->disconnect_all;
builder {
    enable 'AccessLog', logger => sub { };
    sub {
        my $env = shift;
        my $req = Plack::Request->new($env);
        my $id = $req->parameters->{id};
        my $counter = $mem->incr('counter:'.$id);
        [200, ['Content-Type'=>'application/json; charset=UTF-8'], [$_JSON->encode({counter=>$counter})]];
    }
};
```

### Gazelle

```
[ec2-user@ip- ~]$ start_server --backlog=16384 --path=/dev/shm/app.sock  -- plackup -s Gazelle    --max-workers 16 --max-reqs-per-child 500000 -E prod    -a app.psgi

[ec2-user@ip- wrk]$ ./wrk -t 4 -c 500 -d 30 'http://localhost/?id=1'
Running 30s test @ http://localhost/?id=1
  4 threads and 500 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency     5.51ms    6.34ms 178.38ms   99.57%
    Req/Sec    24.27k     1.67k   27.08k    87.92%
  2897691 requests in 30.00s, 574.66MB read
Requests/sec:  96577.25
Transfer/sec:     19.15MB
```

### Starlet

```
[ec2-user@ip- ~]$ start_server --backlog=16384 --path=/dev/shm/app.sock  -- plackup -s Starlet    --max-workers 16 --max-reqs-per-child 500000 -E prod    -a app.psgi

[ec2-user@ip- wrk]$ ./wrk -t 4 -c 500 -d 30 'http://localhost/?id=1'
Running 30s test @ http://localhost/?id=1
  4 threads and 500 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency     9.42ms   16.46ms 352.30ms   99.26%
    Req/Sec    15.22k     1.44k   17.45k    89.50%
  1817375 requests in 30.01s, 359.36MB read
Requests/sec:  60561.77
Transfer/sec:     11.98MB
```

### starman

```
[ec2-user@ip- ~]$ starman --listen /dev/shm/app.sock --backlog=16384 --workers 16 --max-requests 500000 -E prod --preload-app   --disable-keepalive -a app.psgi

[ec2-user@ip- wrk]$ ./wrk -t 4 -c 500 -d 30 'http://localhost/?id=1'
Running 30s test @ http://localhost/?id=1
  4 threads and 500 connections
t2  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency    13.06ms   18.95ms 385.51ms   98.31%
    Req/Sec    10.96k     2.09k   13.84k    71.33%
  1307919 requests in 30.01s, 258.32MB read
Requests/sec:  43588.36
Transfer/sec:      8.61MB
```