Simple web server benchmark - vibe.d is slower than node.js and Go?

[Arun Chandrasekaran]

On Friday, 11 May 2018 at 07:56:04 UTC, Daniel Kozak wrote:
> On Wednesday, 9 May 2018 at 22:37:22 UTC, Arun Chandrasekaran 
> wrote:
>> [...]
>
> I have change my example a little:
>
> case "/": res.writeBody("Hello World " ~ 
> to!string(thisThreadID), "text/plain");
>
> And I get this (siege -p -c15 0b -t 10s http://127.0.0.1:3000 | 
> grep World | sort | uniq):
>
> Hello World 140064214951680
> Hello World 140064223344384
> Hello World 140064231737088
> Hello World 140064240129792
> Hello World 140064248522496
> Hello World 140064256915200
>
> Si I get six different thread ids, which is OK because I have 6 
> cores

siege makes a difference. Earlier I had two chrome windows open 
and I just tried a simple GET from each, with a considerable 
delay and I saw the same thread ID in both the responses. Thanks 
for your help! But couldn't understand how two chrome windows 
could consistently get the same thread ID. netstat shows one 
ESTABLISHED socket. May be Chrome multiplexes the connection? 
That doesn't sound right to me.