Issue #14492 has been updated by shevegen (Robert A. Heiler).


> So not such a big difference.

I also ran it just because I was curious.

> time ruby -e exit

real	0m0.109s
user	0m0.097s
sys	0m0.012s

> time ruby --disable-gems -e exit

real	0m0.015s
user	0m0.005s
sys	0m0.010s

It is not a huge difference but almost noticable. Improvements there
are probably useful simply because most ruby hackers may use gems and
the gem ecosystem, so they may benefit. May not only be old hardware
but also smaller systems where you may still use ruby.

It also somewhat fits to matz' theme/goal to make ruby faster, even
if it may be a tiny benefit compared to e. g. gains achieved via
mjit. Many tiny baby steps versus one big giant step - well, ideally
both kind of steps together ... :)

----------------------------------------
Feature #14492: iseq loading + caching should be in core
https://bugs.ruby-lang.org/issues/14492#change-70438

* Author: normalperson (Eric Wong)
* Status: Open
* Priority: Normal
* Assignee: ko1 (Koichi Sasada)
* Target version: 
----------------------------------------
```
... And not in a RubyGem like yomikomu(*)

RubyGems itself is responsible for over 100ms of startup time on my system(**),
so it would be beneficial to make it part of core and speed up rubygems (and stdlib).
$ time ruby -e exit

real    0m0.160s
user    0m0.155s
sys     0m0.004s

$ time ruby --disable=gems -e exit

real    0m0.014s
user    0m0.013s
sys     0m0.000s

(*) git clone https://github.com/ko1/yomikomu.git

(**) I admit, I am intent on continuing use of Ruby on a laptop from 2005.
     That's roughly when I started using Ruby, so any hardware which Ruby
     worked well on back then should work equally well for current and future
     versions of Ruby.

     And maybe this summer I'll dig out a 600 MHz Duron from the early 2000s :D
```




-- 
https://bugs.ruby-lang.org/

Unsubscribe: <mailto:ruby-core-request / ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>