Jeremy Evans 3de20efc30 Avoid unnecessary array allocations for f(arg, *arg, **arg, **arg), f(*arg, a: lvar), and other calls
The `f(arg, *arg, **arg, **arg)` case was previously not optimized.
The optimizer didn't optimize this case because of the multiple
keyword splats, and the compiler didn't optimize it because the
`f(*arg, **arg, **arg)` optimization added in
0ee3960685e283d8e75149a8777eb0109d41509a didn't apply.

I found it difficult to apply this optimization without changing
the `setup_args_core` API, since by the time you get to the ARGSCAT
case, you don't know whether you were called recursively or directly,
so I'm not sure if it was possible to know at that point whether the
array allocation could be avoided.

This changes the dup_rest argument in `setup_args_core` from an int
to a pointer to int.  This allows us to track whether we have allocated
a caller side array for multiple splats or splat+post across
recursive calls. Check the pointed value (*dup_rest) to determine the
`splatarray` argument. If dup_rest is 1, then use `splatarray true`
(caller-side array allocation), then set *dup_rest back to 0, ensuring
only a single `splatarray true` per method call.

Before calling `setup_args_core`, check whether the array allocation
can be avoided safely using `splatarray false`.  Optimizable cases are:

```
// f(*arg)
SPLAT

// f(1, *arg)
ARGSCAT
 LIST

// f(*arg, **arg)
ARGSPUSH
 SPLAT
 HASH nd_brace=0

// f(1, *arg, **arg)
ARGSPUSH
  ARGSCAT
   LIST
  HASH nd_brace=0
```

If so, dup_rest is set to 0 instead of 1 to avoid the allocation.

After calling `setup_args_core`, check the flag. If the flag
includes `VM_CALL_ARGS_SPLAT`, and the pointed value has changed,
indicating `splatarray true` was used, then also set
`VM_CALL_ARGS_SPLAT_MUT` in the flag.

My initial attempt at this broke the `f(*ary, &ary.pop)` test,
because we were not duplicating the ary in the splat even though
it was modified later (evaluation order issue). The initial attempt
would also break `f(*ary, **ary.pop)` or `f(*ary, kw: ary.pop)` cases
for the same reason. I added test cases for those evaluation
order issues.

Add setup_args_dup_rest_p static function that checks that a given
node is safe.  Call that on the block pass node to determine if
the block pass node is safe.  Also call it on each of the hash
key/value nodes to test that they are safe.  If any are not safe,
then set dup_rest = 1 so that `splatarray true` will be used to
avoid the evaluation order issue.

This new approach has the affect of optimizing most cases of
literal keywords after positional splats.  Previously, only
static keyword hashes after positional splats avoided array
allocation for the splat.  Now, most dynamic keyword hashes
after positional splats also avoid array allocation.

Add allocation tests for dynamic keyword keyword hashes after
positional splats.

setup_args_dup_rest_p is currently fairly conservative. It
could definitely be expanded to handle additional node types
to reduce allocations in additional cases.
2024-07-18 22:17:21 -07:00
2024-07-08 12:24:33 +02:00
2024-05-02 10:12:58 -04:00
2024-07-15 21:08:13 -04:00
2024-07-18 19:34:10 +09:00
2024-06-04 12:40:08 +09:00
2024-07-09 10:22:29 -07:00
2024-07-03 12:32:53 -07:00
2024-04-16 17:20:35 +02:00
2024-05-18 15:44:24 +09:00
2024-04-24 15:30:43 -04:00
2024-04-15 12:08:07 +09:00
2024-04-07 16:44:42 +09:00
2024-06-26 13:31:40 -04:00
2024-04-07 17:29:23 +09:00
2024-07-18 11:20:17 +09:00
2024-07-12 14:43:33 +01:00
GPL
2024-07-08 12:24:33 +02:00
2024-07-08 12:24:33 +02:00
2024-04-27 21:55:28 +09:00
2024-07-08 12:24:33 +02:00
2024-06-18 09:28:25 -07:00
2024-03-19 10:59:25 -07:00
2024-04-04 13:17:26 +09:00
2024-04-29 21:26:44 +09:00
2024-07-09 09:43:07 -04:00
2024-04-15 12:08:07 +09:00
2024-04-07 21:16:26 +09:00
2024-04-04 18:20:13 +09:00
2024-05-18 15:44:24 +09:00
2024-04-16 17:20:35 +02:00
2024-04-08 11:13:29 +09:00
2024-07-09 06:43:28 +09:00
2024-04-15 12:08:07 +09:00
2024-07-05 21:20:54 +09:00
2024-06-18 09:28:25 -07:00
2024-05-06 23:56:01 +09:00

Actions Status: MinGW Actions Status: RJIT Actions Status: Ubuntu Actions Status: Windows Travis Status

What is Ruby?

Ruby is an interpreted object-oriented programming language often used for web development. It also offers many scripting features to process plain text and serialized files, or manage system tasks. It is simple, straightforward, and extensible.

Features of Ruby

  • Simple Syntax
  • Normal Object-oriented Features (e.g. class, method calls)
  • Advanced Object-oriented Features (e.g. mix-in, singleton-method)
  • Operator Overloading
  • Exception Handling
  • Iterators and Closures
  • Garbage Collection
  • Dynamic Loading of Object Files (on some architectures)
  • Highly Portable (works on many Unix-like/POSIX compatible platforms as well as Windows, macOS, etc.) cf. https://docs.ruby-lang.org/en/master/maintainers_md.html#label-Platform+Maintainers

How to get Ruby

For a complete list of ways to install Ruby, including using third-party tools like rvm, see:

https://www.ruby-lang.org/en/downloads/

You can download release packages and the snapshot of the repository. If you want to download whole versions of Ruby, please visit https://www.ruby-lang.org/en/downloads/releases/.

Download with Git

The mirror of the Ruby source tree can be checked out with the following command:

$ git clone https://github.com/ruby/ruby.git

There are some other branches under development. Try the following command to see the list of branches:

$ git ls-remote https://github.com/ruby/ruby.git

You may also want to use https://git.ruby-lang.org/ruby.git (actual master of Ruby source) if you are a committer.

How to build

See Building Ruby

Ruby home page

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

Documentation

Mailing list

There is a mailing list to discuss Ruby. To subscribe to this list, please send the following phrase:

join

in the mail subject (not body) to the address ruby-talk-request@ml.ruby-lang.org.

Copying

See the file COPYING.

Feedback

Questions about the Ruby language can be asked on the Ruby-Talk mailing list or on websites like https://stackoverflow.com.

Bugs should be reported at https://bugs.ruby-lang.org. Read "Reporting Issues" for more information.

Contributing

See "Contributing to Ruby", which includes setup and build instructions.

The Author

Ruby was originally designed and developed by Yukihiro Matsumoto (Matz) in 1995.

matz@ruby-lang.org

Description
Languages
Ruby 58.9%
C 29.5%
Rust 6.1%
C++ 2.9%
Yacc 0.9%
Other 1.6%