Why not show the multiple-threads hunchentoot benchmark ? it's really fast.

[gihnius]

On my MBP( i7cpu 8core), it's over 22k requests/s

$ wrk -c 100 -t 4 -d 10 http://127.0.0.1:5000                                                              
Running 10s test @ http://127.0.0.1:5000
  4 threads and 100 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency    13.22ms   84.65ms   1.72s    97.82%
    Req/Sec     5.91k     1.90k   19.82k    77.95%
  226515 requests in 10.03s, 33.92MB read
  Socket errors: connect 0, read 135, write 0, timeout 20
Requests/sec:  22573.23
Transfer/sec:      3.38MB

#10 threads
$ wrk -c 100 -t 10 -d 10 http://127.0.0.1:5000                                                             
Running 10s test @ http://127.0.0.1:5000
  10 threads and 100 connections
  Thread Stats   Avg      Stdev     Max   +/- Stdev
    Latency    16.16ms   98.85ms   1.95s    97.60%
    Req/Sec     2.48k     1.17k   10.52k    73.37%
  229549 requests in 10.05s, 34.37MB read
  Socket errors: connect 0, read 151, write 0, timeout 15
Requests/sec:  22851.01
Transfer/sec:      3.42MB

This will make others surprised!

[fukamachi]

Because when I did it before, it was slower than even single-threaded Node.js. However, it would be possible to add it.

[digikar99]

Anyone tried setting max-thread-count and max-accept-count on hunchentoot's default one-thread-per-connection-taskmaster?