Jean Boussier <[email protected]> wrote:
> > I guess... I was wondering more in terms of the big picture with
> > HTTP parsing, I/O, and response generation taken into account.
> 
> I wasn't too sure how to benchmark the entire loop. But regardless
> 18/20M i/s shows that it should be meaningless. It's only executed
> once per request, so would "waste" 1 second for 18 million requests
> processed.

Yeah, it can't support persistent connections, so Unix sockets
are required to avoid TCP port exhaustion and ensure repeatable
results.

> > (and there may be improvements in those areas later this year,
> > assuming the world doesn't end sooner...)
> 
> If you are interested in general optimizations, I did spot a bunch
> of sub optimal patterns, such as `=~` in places where a string
> comparison, or `match?` would do.
> 
> However many optimizations are only available on more recent
> rubies.

`match?' is Ruby 2.4+, which is probably too big a jump since
we're still on Ruby 1.9.3 at the moment... (though maybe 2.3+
is/was on the horizon).

String comparison as in `==' and `!='?  Would be interested to know
where and what improvements can be had.

But yeah, Ruby just seems hopeless performance-wise (and
compatibility-wise :P); so unicorn mainly exists to keep legacy
projects alive.
--
unsubscribe: one-click, see List-Unsubscribe header
archive: https://yhbt.net/unicorn-public/

Reply via email to