feat: core lib improvments & multiple bug fixes #331

Chronostasys · 2023-09-09T08:46:52Z

This is a large pr. As porting libc and improve std lib are long term tasks, I decided to
merge this pr before its done as there's already too much bug fixes and code changes.

modification list

trait refactors & bug fixes
- fix many bugs when combining trait with generics
- enhance trait cast ability
improve std lib & core lib
- port more libc functions & provide some abstractions
- add hash function for all primitive types and most of core types
- add hashtable structure
gc improvements
- fix global tracing bugs
- fix large obj madvice not aligned ub
- fix wrong finding immix block hole logic & rm useless variables
lsp improvements
- add basic codelens support

close #340
close #321
half done #316

github-actions · 2023-09-09T08:53:33Z

Benchmark for `7898a07`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	81.1±89.12ns	86.3±88.87ns	+6.41%
allocation bench/singlethread gc alloc benchmark small objects	14.1±5.27ns	13.5±4.26ns	-4.26%
multithreads(2) gc benchmark--65535 small objects(per thread)	1158.2±60.59µs	1211.3±245.10µs	+4.58%
plimmixgc/multi-thread gc stress benchmark small objects	933.7±41.87ms	924.6±46.00ms	-0.97%
plimmixgc/singlethread gc stress benchmark small objects	807.6±18.54ms	805.0±15.23ms	-0.32%
singlethread gc benchmark--65535 small objects	1092.9±2.70µs	1083.8±2.25µs	-0.83%

github-actions · 2023-09-11T03:10:57Z

Benchmark for `1873461`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	79.4±90.17ns	81.1±92.47ns	+2.14%
allocation bench/singlethread gc alloc benchmark small objects	14.9±0.57ns	14.9±1.04ns	0.00%
multithreads(2) gc benchmark--65535 small objects(per thread)	1263.7±269.87µs	1231.9±281.37µs	-2.52%
plimmixgc/multi-thread gc stress benchmark small objects	1032.3±50.74ms	1032.1±60.36ms	-0.02%
plimmixgc/singlethread gc stress benchmark small objects	927.9±19.56ms	925.6±20.02ms	-0.25%
singlethread gc benchmark--65535 small objects	1119.2±5.78µs	1103.8±1.77µs	-1.38%

1. add monad functions to `Result` type 2. update io read function to use `Result` monad 3. fix some of the generated visit function not found during link time when optimization level is greater than 0. 4. update test expects

github-actions · 2023-09-11T09:41:27Z

Benchmark for `edb083a`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	77.9±78.13ns	79.2±64.42ns	+1.67%
allocation bench/singlethread gc alloc benchmark small objects	15.2±2.99ns	15.3±2.43ns	+0.66%
multithreads(2) gc benchmark--65535 small objects(per thread)	1328.6±378.26µs	1317.4±271.08µs	-0.84%
plimmixgc/multi-thread gc stress benchmark small objects	1205.5±26.76ms	1168.3±28.90ms	-3.09%
plimmixgc/singlethread gc stress benchmark small objects	1081.7±19.02ms	1086.1±23.19ms	+0.41%
singlethread gc benchmark--65535 small objects	1132.6±10.83µs	1103.1±7.47µs	-2.60%

codecov · 2023-09-11T09:41:50Z

Codecov Report

Merging #331 (7762482) into master (5e6ff62) will increase coverage by 0.58%.
The diff coverage is 81.15%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #331      +/-   ##
==========================================
+ Coverage   85.80%   86.38%   +0.58%     
==========================================
  Files         100      103       +3     
  Lines       21214    22126     +912     
==========================================
+ Hits        18202    19114     +912     
  Misses       3012     3012

Files	Coverage Δ
immix/src/allocator/global_allocator.rs	`77.44% <100.00%> (ø)`
immix/src/bigobj.rs	`75.00% <100.00%> (ø)`
immix/src/block.rs	`92.53% <100.00%> (-0.09%)`	⬇️
immix/src/mmap.rs	`76.92% <100.00%> (+2.68%)`	⬆️
src/ast/builder/mod.rs	`33.33% <ø> (ø)`
src/ast/diag.rs	`61.65% <ø> (ø)`
src/ast/node/cast.rs	`94.93% <100.00%> (ø)`
src/ast/node/function/generator.rs	`98.94% <100.00%> (+<0.01%)`	⬆️
src/ast/node/implement.rs	`90.71% <100.00%> (+0.67%)`	⬆️
src/ast/node/interface.rs	`85.95% <100.00%> (-3.04%)`	⬇️
... and 33 more

... and 7 files with indirect coverage changes

github-actions · 2023-09-11T10:05:40Z

Benchmark for `98c6698`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	75.7±69.51ns	77.7±62.31ns	+2.64%
allocation bench/singlethread gc alloc benchmark small objects	15.1±1.51ns	16.4±7.55ns	+8.61%
multithreads(2) gc benchmark--65535 small objects(per thread)	1217.4±162.80µs	1138.9±50.14µs	-6.45%
plimmixgc/multi-thread gc stress benchmark small objects	1035.5±27.41ms	1016.4±28.26ms	-1.84%
plimmixgc/singlethread gc stress benchmark small objects	934.1±17.37ms	932.8±23.09ms	-0.14%
singlethread gc benchmark--65535 small objects	1118.9±3.39µs	1092.1±3.39µs	-2.40%

github-actions · 2023-09-12T01:24:57Z

Benchmark for `88d1804`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	86.1±93.24ns	91.8±120.51ns	+6.62%
allocation bench/singlethread gc alloc benchmark small objects	17.8±1.80ns	17.9±2.65ns	+0.56%
multithreads(2) gc benchmark--65535 small objects(per thread)	1532.1±70.34µs	1449.4±67.81µs	-5.40%
plimmixgc/multi-thread gc stress benchmark small objects	1186.8±28.44ms	1199.0±51.23ms	+1.03%
plimmixgc/singlethread gc stress benchmark small objects	1090.1±23.10ms	1104.7±20.35ms	+1.34%
singlethread gc benchmark--65535 small objects	1391.8±42.12µs	1341.5±30.87µs	-3.61%

module之前包含的所有submodule都是直接存的值，每次clone整个依赖树跟着clone，文件一多直接卡爆现在我把submodule搞了个rc，直接300ms->20ms，遥遥领先！

github-actions · 2023-09-12T08:36:21Z

Benchmark for `c3dfb7c`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	70.5±41.68ns	76.6±44.78ns	+8.65%
allocation bench/singlethread gc alloc benchmark small objects	20.0±6.85ns	18.4±1.31ns	-8.00%
multithreads(2) gc benchmark--65535 small objects(per thread)	1482.4±288.75µs	1559.5±254.93µs	+5.20%
plimmixgc/multi-thread gc stress benchmark small objects	1408.2±81.29ms	1349.3±36.64ms	-4.18%
plimmixgc/singlethread gc stress benchmark small objects	1226.9±45.82ms	1208.9±50.17ms	-1.47%
singlethread gc benchmark--65535 small objects	1251.5±128.91µs	1271.1±98.73µs	+1.57%

github-actions · 2023-09-13T07:48:55Z

Benchmark for `d096712`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	71.7±38.96ns	85.1±95.25ns	+18.69%
allocation bench/singlethread gc alloc benchmark small objects	18.9±7.46ns	18.1±3.76ns	-4.23%
multithreads(2) gc benchmark--65535 small objects(per thread)	1526.6±381.28µs	1495.7±329.50µs	-2.02%
plimmixgc/multi-thread gc stress benchmark small objects	1216.7±67.51ms	1192.7±30.89ms	-1.97%
plimmixgc/singlethread gc stress benchmark small objects	1080.9±22.07ms	1064.5±17.25ms	-1.52%
singlethread gc benchmark--65535 small objects	1333.7±19.74µs	1304.3±14.96µs	-2.20%

github-actions · 2023-09-13T08:42:50Z

Benchmark for `ce43835`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	80.4±85.52ns	85.8±101.39ns	+6.72%
allocation bench/singlethread gc alloc benchmark small objects	13.7±3.18ns	14.0±4.23ns	+2.19%
multithreads(2) gc benchmark--65535 small objects(per thread)	1239.6±273.12µs	1194.9±257.63µs	-3.61%
plimmixgc/multi-thread gc stress benchmark small objects	912.1±24.72ms	913.9±18.88ms	+0.20%
plimmixgc/singlethread gc stress benchmark small objects	802.3±17.96ms	831.7±19.33ms	+3.66%
singlethread gc benchmark--65535 small objects	1088.6±2.53µs	1100.1±4.24µs	+1.06%

1. generic trait implicit cast use placeholder function to build vtable 2. array visit function's counter noy initialized to zero

github-actions · 2023-09-15T09:51:25Z

Benchmark for `1c9988c`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	77.1±77.03ns	99.5±111.28ns	+29.05%
allocation bench/singlethread gc alloc benchmark small objects	15.2±2.05ns	15.1±1.57ns	-0.66%
multithreads(2) gc benchmark--65535 small objects(per thread)	1409.5±66.52µs	1405.8±225.42µs	-0.26%
plimmixgc/multi-thread gc stress benchmark small objects	1064.8±24.88ms	1088.8±62.07ms	+2.25%
plimmixgc/singlethread gc stress benchmark small objects	966.6±23.96ms	959.6±19.38ms	-0.72%
singlethread gc benchmark--65535 small objects	1215.1±49.57µs	1161.6±22.32µs	-4.40%

github-actions · 2023-09-15T10:20:53Z

Benchmark for `25a8ce7`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	73.5±36.13ns	78.9±80.00ns	+7.35%
allocation bench/singlethread gc alloc benchmark small objects	20.4±15.61ns	19.8±10.57ns	-2.94%
multithreads(2) gc benchmark--65535 small objects(per thread)	1540.5±402.55µs	1592.3±132.26µs	+3.36%
plimmixgc/multi-thread gc stress benchmark small objects	1260.6±60.30ms	1240.6±56.68ms	-1.59%
plimmixgc/singlethread gc stress benchmark small objects	1114.3±26.43ms	1098.8±18.14ms	-1.39%
singlethread gc benchmark--65535 small objects	1353.1±26.37µs	1338.0±17.36µs	-1.12%

Previous implementation set the generic type to it's trait bound type during analysis, witch causing issue during code gen process, also generation useless code even when a generic function is not use. This commit will construct a new temporary type which auto implements the corrsponding trait bounds, and avoid code gen completely when generic function is not instanciated.

github-actions · 2023-09-15T10:28:20Z

Benchmark for `ccbb9d0`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	73.6±45.83ns	88.5±102.18ns	+20.24%
allocation bench/singlethread gc alloc benchmark small objects	18.2±3.94ns	18.2±4.06ns	0.00%
multithreads(2) gc benchmark--65535 small objects(per thread)	1470.6±265.73µs	1538.3±339.87µs	+4.60%
plimmixgc/multi-thread gc stress benchmark small objects	1207.5±53.27ms	1187.0±36.34ms	-1.70%
plimmixgc/singlethread gc stress benchmark small objects	1051.9±27.17ms	1056.7±31.09ms	+0.46%
singlethread gc benchmark--65535 small objects	1358.7±10.11µs	1358.7±3.05µs	0.00%

github-actions · 2023-09-22T07:51:14Z

Benchmark for `da21b6d`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	82.9±81.36ns	98.4±128.69ns	+18.70%
allocation bench/singlethread gc alloc benchmark small objects	13.9±6.70ns	14.0±0.79ns	+0.72%
multithreads(2) gc benchmark--65535 small objects(per thread)	1150.2±28.05µs	1138.4±83.09µs	-1.03%
plimmixgc/multi-thread gc stress benchmark small objects	927.7±43.00ms	916.8±29.13ms	-1.17%
plimmixgc/singlethread gc stress benchmark small objects	798.4±20.21ms	804.0±19.66ms	+0.70%
singlethread gc benchmark--65535 small objects	1090.5±8.00µs	1049.2±10.25µs	-3.79%

this may cause segment fault occasionally

github-actions · 2023-09-22T09:38:40Z

Benchmark for `3dbd9c0`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	75.9±68.96ns	96.0±105.36ns	+26.48%
allocation bench/singlethread gc alloc benchmark small objects	18.2±5.44ns	15.0±1.93ns	-17.58%
multithreads(2) gc benchmark--65535 small objects(per thread)	1178.3±215.18µs	1192.7±259.33µs	+1.22%
plimmixgc/multi-thread gc stress benchmark small objects	1206.6±100.70ms	1225.4±52.74ms	+1.56%
plimmixgc/singlethread gc stress benchmark small objects	972.1±44.61ms	930.1±38.50ms	-4.32%
singlethread gc benchmark--65535 small objects	1038.3±74.17µs	1070.2±89.19µs	+3.07%

github-actions · 2023-09-26T03:58:53Z

Benchmark for `209f776`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	89.5±96.61ns	84.7±90.47ns	-5.36%
allocation bench/singlethread gc alloc benchmark small objects	13.5±2.76ns	14.2±5.72ns	+5.19%
multithreads(2) gc benchmark--65535 small objects(per thread)	1184.0±188.63µs	1204.0±263.25µs	+1.69%
plimmixgc/multi-thread gc stress benchmark small objects	921.2±31.93ms	963.6±71.32ms	+4.60%
plimmixgc/singlethread gc stress benchmark small objects	825.1±23.38ms	821.6±17.13ms	-0.42%
singlethread gc benchmark--65535 small objects	1072.4±7.07µs	1064.3±13.32µs	-0.76%

github-actions · 2023-09-26T06:46:42Z

Benchmark for `b25da3d`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	85.0±90.94ns	85.0±82.07ns	0.00%
allocation bench/singlethread gc alloc benchmark small objects	13.4±3.90ns	14.8±6.43ns	+10.45%
multithreads(2) gc benchmark--65535 small objects(per thread)	1214.3±55.00µs	1175.8±189.55µs	-3.17%
plimmixgc/multi-thread gc stress benchmark small objects	914.7±20.66ms	940.4±34.34ms	+2.81%
plimmixgc/singlethread gc stress benchmark small objects	783.5±15.53ms	817.6±21.44ms	+4.35%
singlethread gc benchmark--65535 small objects	1093.2±5.54µs	1056.5±1.76µs	-3.36%

github-actions · 2023-09-26T09:40:53Z

Benchmark for `1fe2627`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	76.3±49.20ns	76.9±78.89ns	+0.79%
allocation bench/singlethread gc alloc benchmark small objects	19.8±10.89ns	19.8±5.35ns	0.00%
multithreads(2) gc benchmark--65535 small objects(per thread)	1588.2±497.87µs	1451.5±92.07µs	-8.61%
plimmixgc/multi-thread gc stress benchmark small objects	1231.8±44.06ms	1232.6±46.39ms	+0.06%
plimmixgc/singlethread gc stress benchmark small objects	1093.3±16.48ms	1096.2±25.40ms	+0.27%
singlethread gc benchmark--65535 small objects	1331.3±14.15µs	1328.2±11.13µs	-0.23%

github-actions · 2023-09-27T03:57:38Z

Benchmark for `aa161c0`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	71.7±37.78ns	77.1±50.53ns	+7.53%
allocation bench/singlethread gc alloc benchmark small objects	18.6±5.50ns	19.3±1.21ns	+3.76%
multithreads(2) gc benchmark--65535 small objects(per thread)	1541.5±365.48µs	1537.9±277.43µs	-0.23%
plimmixgc/multi-thread gc stress benchmark small objects	1485.1±169.85ms	1178.3±43.30ms	-20.66%
plimmixgc/singlethread gc stress benchmark small objects	1083.9±20.94ms	1071.9±27.85ms	-1.11%
singlethread gc benchmark--65535 small objects	1384.8±18.38µs	1369.1±4.07µs	-1.13%

github-actions · 2023-09-27T08:40:52Z

Benchmark for `6638e8f`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	88.2±100.00ns	89.2±80.73ns	+1.13%
allocation bench/singlethread gc alloc benchmark small objects	13.9±1.81ns	15.6±3.17ns	+12.23%
multithreads(2) gc benchmark--65535 small objects(per thread)	1091.2±119.96µs	1197.8±140.05µs	+9.77%
plimmixgc/multi-thread gc stress benchmark small objects	1070.0±30.02ms	1189.9±36.85ms	+11.21%
plimmixgc/singlethread gc stress benchmark small objects	836.0±31.16ms	985.7±31.40ms	+17.91%
singlethread gc benchmark--65535 small objects	949.5±40.91µs	994.3±45.34µs	+4.72%

github-actions · 2023-09-27T09:19:43Z

Benchmark for `886ebfe`

Click to view benchmark

Test	Base	PR	%
allocation bench/malloc benchmark small objects	76.1±73.61ns	82.4±78.07ns	+8.28%
allocation bench/singlethread gc alloc benchmark small objects	15.3±2.61ns	16.4±2.24ns	+7.19%
multithreads(2) gc benchmark--65535 small objects(per thread)	1334.9±359.04µs	1312.8±273.57µs	-1.66%
plimmixgc/multi-thread gc stress benchmark small objects	1065.4±36.85ms	1056.8±46.23ms	-0.81%
plimmixgc/singlethread gc stress benchmark small objects	949.8±17.65ms	978.7±24.33ms	+3.04%
singlethread gc benchmark--65535 small objects	1133.9±11.46µs	1138.6±10.67µs	+0.41%

Chronostasys added 2 commits September 7, 2023 16:44

feat: add basic codelens

077f4dd

feat: add some io apis

de5aca8

Chronostasys force-pushed the feat/libc branch from de5aca8 to 4872059 Compare September 9, 2023 08:47

Chronostasys marked this pull request as draft September 9, 2023 08:47

fix: gc limit exception & cargo fmt

7ca9aa0

Chronostasys force-pushed the feat/libc branch from 4872059 to 7ca9aa0 Compare September 11, 2023 03:05

Merge branch 'master' into feat/libc

1954235

fix: win libc different signature

885a2a9

fix: gc scanning external constant

aada2a3

fix: use rc to speed up module clone

1e2496a

module之前包含的所有submodule都是直接存的值，每次clone整个依赖树跟着clone，文件一多直接卡爆现在我把submodule搞了个rc，直接300ms->20ms，遥遥领先！

fix: codelens in wrong pos

1fe1e2f

feat: allow trait cast to sub trait

6a6decf

Chronostasys added 2 commits September 14, 2023 19:46

fix some bugs

1cd6fdd

fix: multiple bugs

0e53e63

1. generic trait implicit cast use placeholder function to build vtable 2. array visit function's counter noy initialized to zero

Chronostasys force-pushed the feat/libc branch from 95e086e to 5c9d65f Compare September 15, 2023 10:15

Chronostasys force-pushed the feat/libc branch from 5c9d65f to a4a7753 Compare September 15, 2023 10:23

fix: wrong debug file location for generic

060553c

fix: rm magic operation in global

79b2660

this may cause segment fault occasionally

Chronostasys added 4 commits September 22, 2023 23:19

fix: windows llvm.used imcompatibal with visitor function name

c325654

fix: big obj madv err

b823d1f

feat: add hash table expand && fix generator bugs

e594db7

fix: ass supports to judge primitives impl traits

b1d0a09

fix: abstract implable with generic

e0839ec

chore: better default import code

7bc2f51

chore: update cli version info

a59d111

fix: allow upcast more types to trait

8ee6982

Chronostasys and others added 3 commits September 27, 2023 17:10

feat: add trait instantiatbility check

1675b16

test: add trait cast test

b8a02e7

Merge branch 'master' into feat/libc

7762482

Chronostasys marked this pull request as ready for review September 27, 2023 09:14

Chronostasys requested review from RINNE-TAN and CjiW September 27, 2023 09:15

This was referenced Sep 27, 2023

add hash functions #321

Closed

trait to subtrait cast #340

Closed

Chronostasys changed the title ~~Feat/libc~~ feat: core lib improvments & multiple bug fixes Sep 27, 2023

Chronostasys added this to the 1.0.0 milestone Sep 27, 2023

RINNE-TAN approved these changes Sep 27, 2023

View reviewed changes

Chronostasys merged commit b62dc34 into master Sep 27, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: core lib improvments & multiple bug fixes #331

feat: core lib improvments & multiple bug fixes #331

Chronostasys commented Sep 9, 2023 •

edited

Loading

github-actions bot commented Sep 9, 2023

github-actions bot commented Sep 11, 2023

github-actions bot commented Sep 11, 2023

codecov bot commented Sep 11, 2023 •

edited

Loading

github-actions bot commented Sep 11, 2023

github-actions bot commented Sep 12, 2023

github-actions bot commented Sep 12, 2023

github-actions bot commented Sep 13, 2023

github-actions bot commented Sep 13, 2023

github-actions bot commented Sep 15, 2023

github-actions bot commented Sep 15, 2023

github-actions bot commented Sep 15, 2023

github-actions bot commented Sep 22, 2023

github-actions bot commented Sep 22, 2023

github-actions bot commented Sep 26, 2023

github-actions bot commented Sep 26, 2023

github-actions bot commented Sep 26, 2023

github-actions bot commented Sep 27, 2023

github-actions bot commented Sep 27, 2023

github-actions bot commented Sep 27, 2023

feat: core lib improvments & multiple bug fixes #331

feat: core lib improvments & multiple bug fixes #331

Conversation

Chronostasys commented Sep 9, 2023 • edited Loading

modification list

github-actions bot commented Sep 9, 2023

Benchmark for 7898a07

github-actions bot commented Sep 11, 2023

Benchmark for 1873461

github-actions bot commented Sep 11, 2023

Benchmark for edb083a

codecov bot commented Sep 11, 2023 • edited Loading

Codecov Report

github-actions bot commented Sep 11, 2023

Benchmark for 98c6698

github-actions bot commented Sep 12, 2023

Benchmark for 88d1804

github-actions bot commented Sep 12, 2023

Benchmark for c3dfb7c

github-actions bot commented Sep 13, 2023

Benchmark for d096712

github-actions bot commented Sep 13, 2023

Benchmark for ce43835

github-actions bot commented Sep 15, 2023

Benchmark for 1c9988c

github-actions bot commented Sep 15, 2023

Benchmark for 25a8ce7

github-actions bot commented Sep 15, 2023

Benchmark for ccbb9d0

github-actions bot commented Sep 22, 2023

Benchmark for da21b6d

github-actions bot commented Sep 22, 2023

Benchmark for 3dbd9c0

github-actions bot commented Sep 26, 2023

Benchmark for 209f776

github-actions bot commented Sep 26, 2023

Benchmark for b25da3d

github-actions bot commented Sep 26, 2023

Benchmark for 1fe2627

github-actions bot commented Sep 27, 2023

Benchmark for aa161c0

github-actions bot commented Sep 27, 2023

Benchmark for 6638e8f

github-actions bot commented Sep 27, 2023

Benchmark for 886ebfe

Chronostasys commented Sep 9, 2023 •

edited

Loading

Benchmark for `7898a07`

Benchmark for `1873461`

Benchmark for `edb083a`

codecov bot commented Sep 11, 2023 •

edited

Loading

Benchmark for `98c6698`

Benchmark for `88d1804`

Benchmark for `c3dfb7c`

Benchmark for `d096712`

Benchmark for `ce43835`

Benchmark for `1c9988c`

Benchmark for `25a8ce7`

Benchmark for `ccbb9d0`

Benchmark for `da21b6d`

Benchmark for `3dbd9c0`

Benchmark for `209f776`

Benchmark for `b25da3d`

Benchmark for `1fe2627`

Benchmark for `aa161c0`

Benchmark for `6638e8f`

Benchmark for `886ebfe`