ruby-dev (Japanese) list archive (unofficial mirror)
 help / color / mirror / Atom feed
From: "tenderlovemaking (Aaron Patterson) via ruby-dev" <ruby-dev@ml.ruby-lang.org>
To: ruby-dev@ml.ruby-lang.org
Cc: "tenderlovemaking (Aaron Patterson)" <noreply@ruby-lang.org>
Subject: [ruby-dev:52136] [Ruby master Bug#20489] Ractor behavior strange in ruby master
Date: Wed, 08 Jan 2025 23:27:55 +0000 (UTC)	[thread overview]
Message-ID: <redmine.journal-111373.20250108232754.51663@ruby-lang.org> (raw)
In-Reply-To: <redmine.issue-20489.20240514090024.51663@ruby-lang.org>

Issue #20489 has been updated by tenderlovemaking (Aaron Patterson).

Status changed from Assigned to Feedback

I also can't seem to reproduce this:

```
[aaron@tc-lan-adapter ~/g/ruby (master)]$ time ruby -v test.rb 8 8
ruby 3.3.4 (2024-07-16 revision 425e468d25) [arm64-darwin23]
[0...1, 1...2, 2...3, 3...4, 4...5, 5...6, 6...7, 7...8]
test.rb:43: warning: Ractor is experimental, and the behavior may change in future versions of Ruby! Also there are many implementation issues.

________________________________________________________
Executed in    5.97 secs    fish           external
   usr time   28.54 secs   78.00 micros   28.54 secs
   sys time    0.14 secs  736.00 micros    0.14 secs

[aaron@tc-lan-adapter ~/g/ruby (master)]$ chruby ruby-3.4.1
[aaron@tc-lan-adapter ~/g/ruby (master)]$ time ruby -v test.rb 8 8
ruby 3.4.1 (2024-12-25 revision 48d4efcb85) +PRISM [arm64-darwin24]
[0...1, 1...2, 2...3, 3...4, 4...5, 5...6, 6...7, 7...8]
test.rb:43: warning: Ractor is experimental, and the behavior may change in future versions of Ruby! Also there are many implementation issues.

________________________________________________________
Executed in    6.42 secs    fish           external
   usr time   28.45 secs    0.09 millis   28.45 secs
   sys time    0.09 secs    1.33 millis    0.08 secs

[aaron@tc-lan-adapter ~/g/ruby (master)]$ time ./ruby -v test.rb 8 8
ruby 3.5.0dev (2025-01-08T20:42:35Z master 96f23306f0) +PRISM [arm64-darwin24]
[0...1, 1...2, 2...3, 3...4, 4...5, 5...6, 6...7, 7...8]
test.rb:43: warning: Ractor is experimental, and the behavior may change in future versions of Ruby! Also there are many implementation issues.

________________________________________________________
Executed in    6.04 secs    fish           external
   usr time   28.74 secs    0.10 millis   28.74 secs
   sys time    0.10 secs    1.31 millis    0.10 secs

```

----------------------------------------
Bug #20489: Ractor behavior strange in ruby master
https://bugs.ruby-lang.org/issues/20489#change-111373

* Author: nekoyama32767 (Jinsong Yu)
* Status: Feedback
* Assignee: ko1 (Koichi Sasada)
* ruby -v: ruby 3.4.0dev (2024-05-14T01:58:31Z master 9d01f657b3) [x86_64-linux]
* Backport: 3.1: UNKNOWN, 3.2: UNKNOWN, 3.3: UNKNOWN
----------------------------------------
This is a tarai program
Run`./ruby tarai_ractor.rb 2 8` is to use 2 thread to run 8 times tarai function total, that means 4 times tarai for each ractor(thread).

```
GC.disable

def split_len(len, split)
    ret = []
    mod = len % split
    head = 0
    tail = 0
    split.times do |i|
        if head >= len
            break
        end
        k = 0
        if i < mod then k = 1 end
        tail = tail + (len/split) + k
        ret.append(head...tail)
        head = tail
    end
    return ret
end

def ary_split(ary, split)
    return split_len(ary.length,split)
end
def item_check(item)
    if item[0] != nil
        1 + item_check(item[0]) + item_check(item[1])
    else
        1
    end
end


def tarai(x, y, z) =
    x <= y ? y : tarai(tarai(x-1, y, z),
                       tarai(y-1, z, x),
                       tarai(z-1, x, y))
  
times = ARGV[0].to_i
split = ARGV[1].to_i

p split_len(times, split)
split_len(times, split).each.map do |sp|
    Ractor.new (sp) {
        s = _1
        s.each do 
         tarai(13, 7, 0) 
        end
    }
end.each(&:take)
```
The problem is in ruby 3.1.2 and ruby 3.3
`./ruby tarai_ractor.rb 1 1` has simiular execute time with `./ruby tarai_ractor.rb 8 8` because each thread only run 1 time of tarai function, like follow:

ruby 3.1.2p20 (2022-04-12 revision 4491bb740a) [x86_64-linux]:
```
time ruby exp_ractor_tarai.rb 1 1
[0...1]
<internal:ractor>:267: warning: Ractor is experimental, and the behavior may change in future versions of Ruby! Also there are many implementation issues.

real    0m1.442s
user    0m1.429s
sys     0m0.014s

time ruby exp_ractor_tarai.rb 8 8
[0...1, 1...2, 2...3, 3...4, 4...5, 5...6, 6...7, 7...8]
<internal:ractor>:267: warning: Ractor is experimental, and the behavior may change in future versions of Ruby! Also there are many implementation issues.

real    0m1.857s
user    0m13.817s
sys     0m0.041s
```


But in ruby master(ruby 3.4.0dev) 
ruby 3.4.0dev (2024-05-14T01:58:31Z master 9d01f657b3) [x86_64-linux]
1 ractor 1 tarai:
```
time ../ruby exp_ractor_tarai.rb 1 1

`RubyGems' were not loaded.
`error_highlight' was not loaded.
`did_you_mean' was not loaded.
`syntax_suggest' was not loaded.
[0...1]
exp_ractor_tarai.rb:47: warning: Ractor is experimental, and the behavior may change in future versions of Ruby! Also there are many implementation issues.

real    0m1.671s
user    0m1.666s
sys     0m0.005s
```

8 ractor 8 tarai:

```
time ../ruby exp_ractor_tarai.rb 8 8

`RubyGems' were not loaded.
`error_highlight' was not loaded.
`did_you_mean' was not loaded.
`syntax_suggest' was not loaded.
[0...1, 1...2, 2...3, 3...4, 4...5, 5...6, 6...7, 7...8]
exp_ractor_tarai.rb:47: warning: Ractor is experimental, and the behavior may change in future versions of Ruby! Also there are many implementation issues.

real    0m18.408s
user    1m58.659s
sys     0m0.021s
```
And in ruby 3.4.0dev when run `time ../ruby exp_ractor_tarai.rb 16 16` 16 thread should be used in system monitoring while only 8 threads are used.
Ruby 3.3 and Ruby 3.1.2 do not have this problem.


---Files--------------------------------
thead16_16.png (168 KB)
thread16_8.png (165 KB)


-- 
https://bugs.ruby-lang.org/
_______________________________________________
ruby-dev mailing list -- ruby-dev@ml.ruby-lang.org
To unsubscribe send an email to ruby-dev-leave@ml.ruby-lang.org

  parent reply	other threads:[~2025-01-08 23:28 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-05-14  9:00 [ruby-dev:52086] " nekoyama32767 (Jinsong Yu) via ruby-dev
2024-09-20  1:10 ` [ruby-dev:52115] " hsbt (Hiroshi SHIBATA) via ruby-dev
2025-01-08 22:01 ` [ruby-dev:52135] " luke-gru (Luke Gruber) via ruby-dev
2025-01-08 23:27 ` tenderlovemaking (Aaron Patterson) via ruby-dev [this message]
2025-01-09  0:04 ` [ruby-dev:52137] " tenderlovemaking (Aaron Patterson) via ruby-dev
2025-01-09  0:22 ` [ruby-dev:52138] " tenderlovemaking (Aaron Patterson) via ruby-dev
2025-01-09  1:11 ` [ruby-dev:52139] " tenderlovemaking (Aaron Patterson) via ruby-dev
2025-01-09  4:58 ` [ruby-dev:52140] " nekoyama32767 (Jinsong Yu) via ruby-dev
2025-01-09  5:16 ` [ruby-dev:52141] " nekoyama32767 (Jinsong Yu) via ruby-dev
2025-01-09  5:47 ` [ruby-dev:52142] " nekoyama32767 (Jinsong Yu) via ruby-dev
2025-01-09 19:14 ` [ruby-dev:52143] " tenderlovemaking (Aaron Patterson) via ruby-dev
2025-01-10  1:13 ` [ruby-dev:52144] " luke-gru (Luke Gruber) via ruby-dev
2025-01-10  3:47 ` [ruby-dev:52145] " nekoyama32767 (Jinsong Yu) via ruby-dev

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=redmine.journal-111373.20250108232754.51663@ruby-lang.org \
    --to=ruby-dev@ml.ruby-lang.org \
    --cc=noreply@ruby-lang.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).