Граф коммитов

899 Коммитов

Автор SHA1 Сообщение Дата
BurdetteLamar 5df574d26e [DOC] Fix links 2024-09-10 14:04:41 -04:00
Peter Zhu a7167d0cee Fix ceil when ndigits is large
[Bug #20654]

This commit fixes Integer#ceil and Float#ceil when the number is
negative and ndigits is large such that 10**ndigits is a bignum.

Previously, it would return 0 in such cases. However, this would cause
unexpected behaviour such as:

    puts 1.ceil(-5) # => 100000
    puts 1.ceil(-10) # => 10000000000
    puts 1.ceil(-20) # => 0

This commit changes the last result so that it will return
100000000000000000000.
2024-07-30 08:21:28 -04:00
Peter Zhu 3af2a7fbe1 Fix floor when ndigits is large
[Bug #20654]

This commit fixes Integer#floor and Float#floor when the number is
negative and ndigits is large such that 10**ndigits is a bignum.

Previously, it would return 0 in such cases. However, this would cause
unexpected behaviour such as:

    puts -1.floor(-5) # => -100000
    puts -1.floor(-10) # => -10000000000
    puts -1.floor(-20) # => 0

This commit changes the last result so that it will return
-100000000000000000000.
2024-07-30 08:21:28 -04:00
Aaron Patterson 2c1655314a Revert moving things to Ruby
This is slowing down benchmarks on x86, so lets revert it for now.
2024-07-29 14:18:11 -07:00
BurdetteLamar aaa542d894 Doc for some #ceil and #floor 2024-07-26 11:22:18 -04:00
BurdetteLamar 78f1b835a0 Doc for some #ceil and #floor 2024-07-26 11:22:18 -04:00
Burdette Lamar 30b9912bb7
[DOC] Doc for Float#ceil (#11125) 2024-07-09 09:43:07 -04:00
Peter Zhu ab3fa8dece [DOC] Use backticks instead of HTML tags 2024-07-09 08:55:58 -04:00
Peter Zhu 5de6d0b35f [DOC] Fix granularity calculation
The granularity is calculated as `10 ** ndigits.abs` rather than
`ndigits.abs * 10`. For example, if `ndigits` is `-2`, it should be
`10 ** 2 == 100` rather than `2 * 10 == 20`.
2024-07-09 08:55:58 -04:00
BurdetteLamar a57b4340d0 Doc fixes 2024-07-08 14:35:29 -04:00
Burdette Lamar f5dfadf38b
[DOC] Doc for Integer#floor (#11077) 2024-07-03 16:00:00 -04:00
Aaron Patterson f4b313f733 move Integer#downto to Ruby
Speeds up ChunkyPNG.

The interpreter is about 70% faster:

```
before: ruby 3.4.0dev (2024-07-03T15:16:17Z master 786cf9db48) [arm64-darwin23]
after: ruby 3.4.0dev (2024-07-03T15:32:25Z ruby-downto 0b8b744ce2) [arm64-darwin23]

----------  -----------  ----------  ----------  ----------  -------------  ------------
bench       before (ms)  stddev (%)  after (ms)  stddev (%)  after 1st itr  before/after
chunky-png  892.2        0.1         526.3       1.0         1.65           1.70
----------  -----------  ----------  ----------  ----------  -------------  ------------
Legend:
- after 1st itr: ratio of before/after time for the first benchmarking iteration.
- before/after: ratio of before/after time. Higher is better for after. Above 1 represents a speedup.
```

YJIT is 2.5x faster:

```
before: ruby 3.4.0dev (2024-07-03T15:16:17Z master 786cf9db48) +YJIT [arm64-darwin23]
after: ruby 3.4.0dev (2024-07-03T15:32:25Z ruby-downto 0b8b744ce2) +YJIT [arm64-darwin23]

----------  -----------  ----------  ----------  ----------  -------------  ------------
bench       before (ms)  stddev (%)  after (ms)  stddev (%)  after 1st itr  before/after
chunky-png  709.4        0.1         278.8       0.3         2.35           2.54
----------  -----------  ----------  ----------  ----------  -------------  ------------
Legend:
- after 1st itr: ratio of before/after time for the first benchmarking iteration.
- before/after: ratio of before/after time. Higher is better for after. Above 1 represents a speedup.
```
2024-07-03 10:01:41 -07:00
BurdetteLamar 9930647134 Doc for Integer#ceil 2024-06-29 10:33:54 -04:00
Artur dd578cf2f1
[DOC] Enhance `Numeric#nonzero?` doc
Add `zero?` as a related method
2024-04-26 12:33:41 +00:00
Reznov 19f4b06b9d
Reducing the number of divisions in `rb_fix_digits` 2024-04-08 12:51:32 +09:00
Peter Zhu 9ad175c1ee Register rb_fix_to_s_static as global right after creating
If a GC runs right during creating a rb_fix_to_s_static, it may cause
the previous ones to become swept by the GC because they have not been
registered by rb_vm_register_global_object.
2024-03-27 09:39:23 -04:00
tompng 0ff2c7fe6f Faster Integer.sqrt for large bignum
Integer.sqrt uses Newton's method.
This pull request reduces the precision which was unnecessarily high in each calculation step.
2024-03-18 13:52:27 +09:00
Jean Boussier d4f3dcf4df Refactor VM root modules
This `st_table` is used to both mark and pin classes
defined from the C API. But `vm->mark_object_ary` already
does both much more efficiently.

Currently a Ruby process starts with 252 rooted classes,
which uses `7224B` in an `st_table` or `2016B` in an `RArray`.

So a baseline of 5kB saved, but since `mark_object_ary` is
preallocated with `1024` slots but only use `405` of them,
it's a net `7kB` save.

`vm->mark_object_ary` is also being refactored.

Prior to this changes, `mark_object_ary` was a regular `RArray`, but
since this allows for references to be moved, it was marked a second
time from `rb_vm_mark()` to pin these objects.

This has the detrimental effect of marking these references on every
minors even though it's a mostly append only list.

But using a custom TypedData we can save from having to mark
all the references on minor GC runs.

Addtionally, immediate values are now ignored and not appended
to `vm->mark_object_ary` as it's just wasted space.
2024-03-06 15:33:43 -05:00
Nobuyoshi Nakada 3dccb716da
Use `defined?(yield)` and `SIZED_ENUMERATOR`
Prefer built-in features over method calls that may be overridden.
2024-02-17 23:28:00 +09:00
Yusuke Endoh 25d74b9527 Do not include a backtick in error messages and backtraces
[Feature #16495]
2024-02-15 18:42:31 +09:00
Peter Zhu fd87259a26 Replace assert with RUBY_ASSERT in numeric.c
assert does not print the bug report, only the file and line number of
the assertion that failed. RUBY_ASSERT prints the full bug report, which
makes it much easier to debug.
2024-02-12 15:07:47 -05:00
BurdetteLamar 8cad6fc90a Tweak to Integer What's Here 2024-01-06 13:25:17 -05:00
BurdetteLamar 4a44bf8c34 Tweak to Float What's Here 2024-01-06 13:24:48 -05:00
BurdetteLamar 7b615bfdf3 Tweak to Float What's Here 2024-01-06 13:24:48 -05:00
Peter Zhu d96fe5e4f8 [DOC] Fix indentation in Numeric#step 2024-01-06 11:43:57 -05:00
Peter Zhu 0831d960bf [DOC] Fix typo in call-seq for Float#round 2024-01-06 09:34:54 -05:00
Peter Zhu 6934a60ab1 [DOC] Fix alignment of values for Integer#div 2024-01-01 11:14:42 -05:00
Peter Zhu c0481e5818 [DOC] Fix indentation for Integer#div
The line was indented, which caused it to be treated as a code block.
2024-01-01 11:12:54 -05:00
Peter Zhu 606c01727a [DOC] Fix indentation for Numeric#step
The documentation was indented one level too deep, which causes RDoc to
generate it all as a code block.
2023-12-29 13:13:17 -05:00
Peter Zhu 550a49c913 [DOC] Fix documentation for Numeric#eql?
+==+ does not format properly, so we should use <tt>==</tt> instead.
2023-12-26 10:48:48 -05:00
John Hawthorn 0c3593b657 Use free with ruby_dtoa
In ae0ceafb0c0d05cc80623b525070255e3abb34ef ruby_dtoa was switched to
use malloc instead of xmalloc, which means that consumers should be
using free instead of xfree. Otherwise we will artificially shrink
oldmalloc_increase_bytes.
2023-12-07 09:23:02 -08:00
Nobuyoshi Nakada 8e93bf8e1f
[Bug #17037] Improve accuracy of division near precision limits
When dividing near the precision limit of `double`, use Bignum
division to get rid of rounding errors.
2023-11-29 20:16:36 +09:00
Kouhei Yanagita 5cb76754ab [DOC] Fix the argument name in the Numeric#step doc 2023-10-15 13:47:58 +09:00
Kouhei Yanagita b28c1d2c56 [Bug #19926] Fix Range#size for ranges with a Rational endpoint 2023-10-14 11:22:42 +09:00
Nobuyoshi Nakada 50520cc193
[DOC] Missing comment markers 2023-09-27 16:18:05 +09:00
Takashi Kokubun 5b5ae3d9e0
Rewrite Integer#times in Ruby (#8388) 2023-09-07 10:57:52 -07:00
BurdetteLamar 3080cf3dec [DOC] Don't suppress autolinks 2023-08-13 15:47:12 -04:00
Burdette Lamar 6528cf9fcf
[DOC] Fixes for link fragments (#7981) 2023-06-28 09:05:43 -04:00
Peter Zhu b7a26dfa16 Unify error messages of rb_num2ulong and rb_num2ull
The error messages were slightly different due, which causes different
behaviour on 32-bit and 64-bit systems.
2023-05-29 11:55:39 -04:00
Matt Valentine-House 026321c5b9 [Feature #19474] Refactor NEWOBJ macros
NEWOBJ_OF is now our canonical newobj macro. It takes an optional ec
2023-04-06 11:07:16 +01:00
Takashi Kokubun 233ddfac54 Stop exporting symbols for MJIT 2023-03-06 21:59:23 -08:00
BurdetteLamar 3b239d2480 Remove (newly unneeded) remarks about aliases 2023-02-19 14:26:34 -08:00
Nobuyoshi Nakada dd28c55a7c
[Bug#19445] Fix keyword splat in enumerator
Extracted arguments do not have keyword hash to splat.
2023-02-17 10:57:22 +09:00
Takashi Kokubun 344c16eba4
Avoid using a weird syntax for documentation
Following up 465bd972ec.

If the actual implementation still resides in C, it should be
documented in C just like all other places.
2023-01-30 13:18:19 -08:00
Nobuyoshi Nakada 71ce7e1825
[Bug #19335] `Integer#remainder` should respect `#coerce` (#7120)
Also `Numeric#remainder` should.
2023-01-15 13:03:27 +09:00
Kenta Murata 9f2378959e
numeric.c: Fix round_half_even for specific values (#7023)
Handle the integert and the float parts separately in round_half_even
to prevent error occursions in floating point calculation.
2022-12-26 21:02:47 +09:00
S-H-GAMELINKS 1f4f6c9832 Using UNDEF_P macro 2022-11-16 18:58:33 +09:00
Nobuyoshi Nakada 5ccb625fbb
Use `roomof` macro for rounding up divisions 2022-10-14 19:23:25 +09:00
Kouhei Yanagita 803a072630 Improve performance of Integer#ceildiv
This patch is suggested by nobu.

Benchmark result:

```
require 'benchmark'

n = 10 ** 7

Benchmark.bm do |x|
  x.report("Fixnum/Fixnum") { a, b = 5, 2; n.times { a.ceildiv(b) } }
  x.report("Bignum/Bignum") { a, b = 10**100, 10**99 - 1; n.times { a.ceildiv(b) } }
  x.report("Bignum/Fixnum") { a, b = 10**100, 3; n.times { a.ceildiv(b) } }
end
```

Original:

```
       user     system      total        real
Fixnum/Fixnum  3.340009   0.043029   3.383038 (  3.384022)
Bignum/Bignum  8.229500   0.118543   8.348043 (  8.349574)
Bignum/Fixnum  8.328971   0.097842   8.426813 (  8.426952)
```

Improved:

```
       user     system      total        real
Fixnum/Fixnum  0.699140   0.000961   0.700101 (  0.700199)
Bignum/Bignum  5.076165   0.083160   5.159325 (  5.159360)
Bignum/Fixnum  5.548684   0.115372   5.664056 (  5.666735)
```
2022-08-12 15:57:52 +09:00
Kouhei Yanagita 24e33b84b5 Remove Numeric#ceildiv 2022-08-12 15:57:52 +09:00