YJIT Benchmarks

Details for Benchmarks at 2024-03-05 06:08:14 UTC

YJIT metrics from the yjit-bench suite using Ruby 0ae0a0c1c3

Using the geomean of the headline benchmarks for x86 YJIT 3.4.0dev is
  • 71.4% faster than CRuby 3.4.0dev
On railsbench it is
  • 77.8% faster than CRuby 3.4.0dev

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 3.4.0dev warmups CRuby 3.4.0dev iters YJIT 3.4.0dev warmups YJIT 3.4.0dev iters
activerecord 30 512 30 512
chunky-png 30 32 30 32
erubi-rails 30 19 30 19
hexapdf 30 15 30 15
liquid-c 30 349 30 349
liquid-compile 30 342 30 342
liquid-render 30 254 30 254
lobsters 30 23 30 23
mail 30 162 30 162
psych-load 30 15 30 15
railsbench 30 15 30 15
rubocop 30 181 30 181
ruby-lsp 30 200 30 200
sequel 30 288 30 288
binarytrees 30 84 30 84
blurhash 30 99 30 99
erubi 30 76 30 76
etanni 30 51 30 51
fannkuchredux 30 23 30 23
fluentd 30 15 30 15
graphql 30 15 30 15
graphql-native 30 22 30 22
lee 30 20 30 20
matmul 30 17 30 17
nbody 30 314 30 314
nqueens 30 294 30 294
optcarrot 30 15 30 15
rack 30 179 30 179
ruby-json 30 15 30 15
rubykon 30 15 30 15
sudoku 30 24 30 24
tinygql 30 41 30 41
30k_ifelse 30 55 30 55
30k_methods 30 23 30 23
cfunc_itself 30 827 30 827
fib 30 416 30 416
getivar 30 940 30 940
keyword_args 30 591 30 591
respond_to 30 1480 30 1480
ruby-xor 30 112 30 112
setivar 30 2392 30 2392
setivar_object 30 488 30 488
setivar_young 30 505 30 505
str_concat 30 405 30 405
throw 30 747 30 747

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) 885433 775458 49 532 0 0% 0 0 15.744322
chunky-png (click) 337170 299401 87 1096 1 0% 0 0 32.53955
erubi-rails (click) 1713454 1459443 280 2964 6 0% 0 0 85.445164
hexapdf (click) 1418585 1251624 596 12576 38 0% 0 0 368.638462
liquid-c (click) 525038 456686 119 1754 5 0% 0 0 47.473599
liquid-compile (click) 448043 390540 151 2072 2 0% 0 0 63.489668
liquid-render (click) 623603 549564 136 2283 8 0% 0 0 62.932575
lobsters (click) 7878714 6609124 3217 49934 74 0% 0 0 1708.534077
mail (click) 732395 635954 346 4847 14 0% 0 0 143.372396
psych-load (click) 301215 253619 65 655 3 0% 0 0 20.431394
railsbench (click) 2996392 2572757 1638 15761 54 0% 0 0 454.988342
rubocop (click) 5368130 4570274 2901 47791 106 0% 4 0 1424.650485
ruby-lsp (click) 634011 541577 159 2735 14 0% 0 0 76.770267
sequel (click) 491379 423758 16 120 0 0% 0 0 4.718536
binarytrees (click) 9392 7588 11 75 0 0% 0 0 3.133232
blurhash (click) 52000 47236 32 461 0 0% 0 0 15.42078
erubi (click) 271214 230596 10 111 0 0% 0 0 3.883602
etanni (click) 31686 27819 12 96 0 0% 0 0 3.231244
fannkuchredux (click) 24912 17701 9 257 0 0% 0 0 7.881932
fluentd (click) 387873 311751 11 111 0 0% 0 0 3.608334
graphql (click) 443736 382860 70 673 0 0% 0 0 21.102168
graphql-native (click) 409288 344342 41 271 0 0% 0 0 9.34442
lee (click) 315849 277753 49 731 0 0% 0 0 22.845871
matmul (click) 12586 12175 12 126 0 0% 0 0 4.211258
nbody (click) 15103 7574 11 185 0 0% 0 0 5.032434
nqueens (click) 23609 17866 10 261 0 0% 0 0 7.871977
optcarrot (click) 317087 278432 192 4342 34 0% 0 0 105.111214
rack (click) 272097 234532 31 319 0 0% 0 0 9.201112
ruby-json (click) 28135 23070 12 198 0 0% 0 0 5.826491
rubykon (click) 133625 106471 143 1498 3 0% 0 0 41.809874
sudoku (click) 47007 45195 11 507 0 0% 0 0 15.520241
tinygql (click) 313012 266151 63 788 5 0% 0 0 22.739425
30k_ifelse (click) 5222597 4374236 9264 50813 0 0% 0 0 1482.380916
30k_methods (click) 2019614 1594052 5782 19369 0 0% 0 0 522.044195
cfunc_itself (click) 7589 5810 9 70 0 0% 0 0 2.246761
fib (click) 5017 4633 8 49 0 0% 0 0 1.79611
getivar (click) 6102 6023 8 74 0 0% 0 0 2.136089
keyword_args (click) 8472 6502 10 72 0 0% 0 0 2.539673
respond_to (click) 8341 7760 9 85 0 0% 0 0 2.66536
ruby-xor (click) 7940 7126 9 87 0 0% 0 0 3.028235
setivar (click) 5193 4604 8 54 0 0% 0 0 1.763869
setivar_object (click) 5561 4624 8 54 0 0% 0 0 1.847452
setivar_young (click) 6284 5274 9 62 0 0% 0 0 2.064946
str_concat (click) 7932 7204 11 85 0 0% 0 0 2.843854
throw (click) 7946 6065 10 69 0 0% 0 0 2.492684

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.