YJIT Benchmarks

Details for Benchmarks at 2024-06-22 19:08:15 UTC

YJIT metrics from the yjit-bench suite using Ruby fba8aff7af

Using the geomean of the headline benchmarks for x86 YJIT 3.4.0dev is
  • 71.5% faster than CRuby 3.4.0dev
  • the same speed as YJIT 3.3.3
On railsbench it is
  • 75.9% faster than CRuby 3.4.0dev
  • 1.6% slower than YJIT 3.3.3

Performance on Headline Benchmarks

Speed of each Ruby implementation relative to the baseline CRuby measurement. Higher is better.

Memory Usage on Headline Benchmarks

Memory usage of each Ruby implementation relative to the baseline CRuby measurement. Lower is better.

Performance on Other Benchmarks

Speed of each Ruby implementation relative to the baseline CRuby measurement. Higher is better.

Memory Usage on Other Benchmarks

Memory usage of each Ruby implementation relative to the baseline CRuby measurement. Lower is better.

Performance on MicroBenchmarks

Speed of each Ruby implementation relative to the baseline CRuby measurement. Higher is better.

Memory Usage on MicroBenchmarks

Memory usage of each Ruby implementation relative to the baseline CRuby measurement. Lower is better.

Want Raw Graphs and CSV?

Benchmarks Speed Details

Benchmark Memory Usage Details

Number of Iterations and Warmups Tested

bench CRuby 3.3.3 warmups CRuby 3.3.3 iters CRuby 3.4.0dev warmups CRuby 3.4.0dev iters YJIT 3.3.3 warmups YJIT 3.3.3 iters YJIT 3.4.0dev warmups YJIT 3.4.0dev iters
activerecord 10 51 10 47 10 115 10 116
chunky-png 10 18 10 15 10 36 10 36
erubi-rails 10 10 10 10 10 20 10 19
hexapdf 10 10 10 10 10 10 10 10
liquid-c 10 353 10 333 10 504 10 474
liquid-compile 10 341 10 332 10 490 10 486
liquid-render 10 136 10 133 10 354 10 349
mail 10 162 10 147 10 237 10 218
psych-load 10 10 10 10 10 10 10 10
railsbench 10 10 10 10 10 10 10 10
rubocop 10 137 10 127 10 231 10 235
ruby-lsp 10 182 10 193 10 262 10 281
sequel 10 326 10 303 10 437 10 403
binarytrees 10 51 10 47 10 117 10 113
blurhash 10 61 10 58 10 118 10 143
erubi 10 83 10 83 10 102 10 104
etanni 10 59 10 57 10 68 10 67
fannkuchredux 10 10 10 10 10 18 10 16
fluentd 10 10 10 10 10 10 10 10
graphql 10 10 10 10 10 10 10 10
graphql-native 10 33 10 33 10 39 10 38
lee 10 11 10 10 10 19 10 20
matmul 10 10 10 10 10 15 10 15
nbody 10 211 10 188 10 385 10 467
nqueens 10 115 10 102 10 115 10 425
optcarrot 10 10 10 10 10 10 10 10
protoboeuf 10 189 10 171 10 680 10 658
rack 10 470 10 465 10 809 10 875
ruby-json 10 10 10 10 10 10 10 10
rubykon 10 10 10 10 10 10 10 10
sudoku 10 10 10 10 10 31 10 28
tinygql 10 24 10 22 10 52 10 50
30k_ifelse 10 10 10 10 10 70 10 69
30k_methods 10 10 10 10 10 26 10 26
cfunc_itself 10 257 10 232 10 1105 10 1210
fib 10 103 10 100 10 616 10 618
getivar 10 255 10 242 10 1269 10 1365
keyword_args 10 90 10 83 10 726 10 944
respond_to 10 89 10 83 10 1782 10 2146
ruby-xor 10 31 10 25 10 100 10 127
setivar 10 423 10 406 10 2354 10 3572
setivar_object 10 268 10 265 10 653 10 736
setivar_young 10 264 10 265 10 651 10 728
str_concat 10 327 10 288 10 592 10 506
throw 10 975 10 837 10 1275 10 1037

Different Ruby configurations want different amounts of warmup. With no JIT, CRuby needs hardly any. YJIT and MJIT 3.0 both warm up quite quickly, while MJIT in 3.1 often slows down for a time as it compiles, after an unpredictable delay.

Benchmark YJIT Stats

Hover your cursor over the benchmark names for descriptions of each benchmark.

bench Exit Report Inline Outlined Comp iSeqs Comp Blocks Inval Inval Ratio Bind Alloc Bind Set Const Bumps Compile Time MS
activerecord (click) 1403190 1155146 214 2166 0 0% 0 0 76.055314
chunky-png (click) 331257 268947 82 1053 1 0% 0 0 33.74836
erubi-rails (click) 1677359 1456361 272 2909 6 0% 0 0 96.275728
hexapdf (click) 1288403 1072943 504 11685 41 0% 0 0 373.835105
liquid-c (click) 517841 425381 119 1761 5 0% 0 0 53.296819
liquid-compile (click) 440741 379986 151 2070 2 0% 0 0 69.398471
liquid-render (click) 612809 533640 136 2302 8 0% 0 0 70.410256
mail (click) 723226 644597 346 4881 14 0% 0 0 156.72677
psych-load (click) 293176 243984 61 594 3 0% 0 0 20.592012
railsbench (click) 2924575 2537382 1633 15916 57 0% 0 0 555.223266
rubocop (click) 5432285 4549565 2891 50364 110 0% 4 0 1774.540292
ruby-lsp (click) 632837 548322 159 2903 28 0% 0 0 90.461356
sequel (click) 487839 409390 16 116 0 0% 0 0 5.114378
binarytrees (click) 8756 7777 10 70 0 0% 0 0 2.984003
blurhash (click) 50429 43324 31 457 0 0% 0 0 17.016057
erubi (click) 270214 214579 10 112 0 0% 0 0 4.356563
etanni (click) 30236 28701 12 96 0 0% 0 0 3.486978
fannkuchredux (click) 21198 26809 5 228 0 0% 0 0 7.872639
fluentd (click) 375007 342770 4 28 0 0% 0 0 1.240495
graphql (click) 430898 361977 66 648 0 0% 0 0 22.205978
graphql-native (click) 402926 330733 42 278 0 0% 0 0 10.798638
lee (click) 293365 214005 34 531 0 0% 0 0 18.143055
matmul (click) 8315 983 8 86 0 0% 0 0 2.998869
nbody (click) 14798 17933 11 185 0 0% 0 0 5.514633
nqueens (click) 22145 28339 10 257 0 0% 0 0 8.885101
optcarrot (click) 304463 277860 188 4310 34 0% 0 0 108.908761
protoboeuf (click) 140419 133760 17 1349 0 0% 0 0 43.121378
rack (click) 275741 228622 37 408 0 0% 0 0 11.976556
ruby-json (click) 30222 27890 8 167 0 0% 0 0 5.169692
rubykon (click) 128391 79892 137 1484 3 0% 0 0 45.210782
sudoku (click) 41994 57542 8 473 0 0% 0 0 15.934732
tinygql (click) 307906 259268 58 734 5 0% 0 0 23.191204
30k_ifelse (click) 5134674 4729436 9259 50778 0 0% 0 0 1915.157155
30k_methods (click) 2016920 1591842 5778 19339 0 0% 0 0 531.224622
cfunc_itself (click) 7534 6180 9 70 0 0% 0 0 2.326439
fib (click) 4970 5003 8 49 0 0% 0 0 1.982305
getivar (click) 5963 6375 8 74 0 0% 0 0 2.24096
keyword_args (click) 8409 6912 10 72 0 0% 0 0 2.580415
respond_to (click) 8277 8274 9 85 0 0% 0 0 2.794684
ruby-xor (click) 5565 5673 5 62 0 0% 0 0 1.851106
setivar (click) 5102 5375 8 54 0 0% 0 0 1.939745
setivar_object (click) 5470 5395 8 54 0 0% 0 0 1.998197
setivar_young (click) 6182 6093 9 62 0 0% 0 0 2.210856
str_concat (click) 7774 7946 11 85 0 0% 0 0 3.092672
throw (click) 7895 6377 10 69 0 0% 0 0 2.63493

YJIT stats correspond to the YJIT stats exit report.

Note: currently, all stats are collected on x86_64, not ARM.

Raw JSON data files

All graphs and table data in this page comes from processing these data files, which come from benchmark runs.