[#48745] [ruby-trunk - Bug #7267][Open] Dir.glob on Mac OS X returns unexpected string encodings for unicode file names — "kennygrant (Kenny Grant)" <kennygrant@...>

17 messages 2012/11/02

[#48773] [ruby-trunk - Bug #7269][Open] Refinement doesn't work if using locate after method — "ko1 (Koichi Sasada)" <redmine@...>

12 messages 2012/11/03

[#48847] [ruby-trunk - Bug #7274][Open] UnboundMethods should be bindable to any object that is_a?(owner of the UnboundMethod) — "rits (First Last)" <redmine@...>

21 messages 2012/11/04

[#48854] [ruby-trunk - Bug #7276][Open] TestFile#test_utime failure — "jonforums (Jon Forums)" <redmine@...>

14 messages 2012/11/04

[#48988] [ruby-trunk - Feature #7292][Open] Enumerable#to_h — "marcandre (Marc-Andre Lafortune)" <ruby-core@...>

40 messages 2012/11/06

[#48997] [ruby-trunk - Feature #7297][Open] map_to alias for each_with_object — "nathan.f77 (Nathan Broadbent)" <nathan.f77@...>

19 messages 2012/11/06

[#49001] [ruby-trunk - Bug #7298][Open] Behavior of Enumerator.new different between 1.9.3 and 2.0.0 — "ayumin (Ayumu AIZAWA)" <ayumu.aizawa@...>

12 messages 2012/11/06

[#49018] [ruby-trunk - Feature #7299][Open] Ruby should not completely ignore blocks. — "marcandre (Marc-Andre Lafortune)" <ruby-core@...>

13 messages 2012/11/07

[#49044] [ruby-trunk - Bug #7304][Open] Random test failures around test_autoclose_true_closed_by_finalizer — "luislavena (Luis Lavena)" <luislavena@...>

11 messages 2012/11/07

[#49196] [ruby-trunk - Feature #7322][Open] Add a new operator name #>< for bit-wise "exclusive or" — "alexeymuranov (Alexey Muranov)" <redmine@...>

18 messages 2012/11/10

[#49211] [ruby-trunk - Feature #7328][Open] Move ** operator precedence under unary + and - — "boris_stitnicky (Boris Stitnicky)" <boris@...>

20 messages 2012/11/11

[#49229] [ruby-trunk - Bug #7331][Open] Set the precedence of unary `-` equal to the precedence `-`, same for `+` — "alexeymuranov (Alexey Muranov)" <redmine@...>

17 messages 2012/11/11

[#49256] [ruby-trunk - Feature #7336][Open] Flexiable OPerator Precedence — "trans (Thomas Sawyer)" <transfire@...>

18 messages 2012/11/12

[#49354] review open pull requests on github — Zachary Scott <zachary@...>

Could we get a review on any open pull requests on github before the

12 messages 2012/11/15
[#49355] Re: review open pull requests on github — "NARUSE, Yui" <naruse@...> 2012/11/15

2012/11/15 Zachary Scott <zachary@zacharyscott.net>:

[#49356] Re: review open pull requests on github — Zachary Scott <zachary@...> 2012/11/15

Ok, I was hoping one of the maintainers might want to.

[#49451] [ruby-trunk - Bug #7374][Open] File.expand_path resolving to first file/dir instead of absolute path — mdube@... (Martin Dubé) <mdube@...>

12 messages 2012/11/16

[#49463] [ruby-trunk - Feature #7375][Open] embedding libyaml in psych for Ruby 2.0 — "tenderlovemaking (Aaron Patterson)" <aaron@...>

21 messages 2012/11/16
[#49494] [ruby-trunk - Feature #7375] embedding libyaml in psych for Ruby 2.0 — "vo.x (Vit Ondruch)" <v.ondruch@...> 2012/11/17

[#49467] [ruby-trunk - Feature #7377][Open] #indetical? as an alias for #equal? — "aef (Alexander E. Fischer)" <aef@...>

13 messages 2012/11/17

[#49558] [ruby-trunk - Bug #7395][Open] Negative numbers can't be primes by definition — "zzak (Zachary Scott)" <zachary@...>

10 messages 2012/11/19

[#49566] [ruby-trunk - Feature #7400][Open] Incorporate OpenSSL tests from JRuby. — "zzak (Zachary Scott)" <zachary@...>

11 messages 2012/11/19

[#49770] [ruby-trunk - Feature #7414][Open] Now that const_get supports "Foo::Bar" syntax, so should const_defined?. — "robertgleeson (Robert Gleeson)" <rob@...>

9 messages 2012/11/20

[#49950] [ruby-trunk - Feature #7427][Assigned] Update Rubygems — "mame (Yusuke Endoh)" <mame@...>

17 messages 2012/11/24

[#50043] [ruby-trunk - Bug #7429][Open] Provide options for core collections to customize behavior — "headius (Charles Nutter)" <headius@...>

10 messages 2012/11/24

[#50092] [ruby-trunk - Feature #7434][Open] Allow caller_locations and backtrace_locations to receive negative params — "sam.saffron (Sam Saffron)" <sam.saffron@...>

21 messages 2012/11/25

[#50094] [ruby-trunk - Bug #7436][Open] Allow for a "granularity" flag for backtrace_locations — "sam.saffron (Sam Saffron)" <sam.saffron@...>

11 messages 2012/11/25

[#50207] [ruby-trunk - Bug #7445][Open] strptime('%s %z') doesn't work — "felipec (Felipe Contreras)" <felipe.contreras@...>

19 messages 2012/11/27

[#50424] [ruby-trunk - Bug #7485][Open] ruby cannot build on mingw32 due to missing __sync_val_compare_and_swap — "drbrain (Eric Hodel)" <drbrain@...7.net>

15 messages 2012/11/30

[#50429] [ruby-trunk - Feature #7487][Open] Cutting through the issues with Refinements — "trans (Thomas Sawyer)" <transfire@...>

13 messages 2012/11/30

[ruby-core:49754] [ruby-trunk - Feature #6962][Assigned] Use lighter hash structure for methods table, instance variable positions, constants

From: "mame (Yusuke Endoh)" <mame@...>
Date: 2012-11-20 14:27:12 UTC
List: ruby-core #49754
Issue #6962 has been updated by mame (Yusuke Endoh).

Status changed from Open to Assigned
Assignee changed from matz (Yukihiro Matsumoto) to ko1 (Koichi Sasada)
Target version set to next minor

ko1 or nobu, please check this (towards next minor).

-- 
Yusuke Endoh <mame@tsg.ne.jp>
----------------------------------------
Feature #6962: Use lighter hash structure for methods table, instance variable positions, constants
https://bugs.ruby-lang.org/issues/6962#change-33333

Author: funny_falcon (Yura Sokolov)
Status: Assigned
Priority: Normal
Assignee: ko1 (Koichi Sasada)
Category: core
Target version: next minor


I've experimented with replacing `struct st_table` with lighter hash structure.

I were inspired by Lua tables, so that hash were created similar:
- open addressing but with elements chaining instead of any kind of probing
- there is sibling probing on insert, to ensure cache locality
- on insert, already inserted collision item which is not in main position is moved to free place,
  so that chains started from different main positions are not joined.
  This allows to keep chains short.
- struct header is 24 bytes on 64bit platform (compared with 40 bytes for `st_table`), so that i've
  placed them directly into `rb_classext_struct` instead of separate allocation.

Since keys are integers (ID), there is no separate field for hash. And I used `unsigned int` instead of
`unsigned long` for key, cause ID almost never reach 2**32. It allows to pack whole entry into 16 bytes
on 64bit platform (compared to 48 bytes for `st_table_entry`).

Such structure gives about 6% improvement without GC tuning, and 10% improvement when GC is tuned
(I've used `RUBY_GC_MALLOC_LIMIT=60000000 RUBY_FREE_MIN=200000`)

Also I add per class method cache instead of single global hash array. It has advantage over global
cache cause there is no collision over different classes and methods. But it seems that Rails applications
too often creates singleton classes, so that improvement is not very big (but no slowdown too).

Currently I have patch for Ruby 1.9.3 https://github.com/funny-falcon/ruby/compare/ruby_1_9_3...sparse_array/ruby_1_9_3

Single incompatible external api change is almost neglible, cause `rb_generic_ivar_table` methos is not very usefull
https://github.com/funny-falcon/ruby/compare/ruby_1_9_3...sparse_array/ruby_1_9_3#L5L852

Silly, but there is too many changes in trunk, so that, I could not simply rebase it, but ought to make whole work again.
I'll try to do it shortly.

With regards, Sokolov Yura aka funny-falcon.


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

In This Thread