YJIT Benchmarks

Details for Benchmarks at 2022-04-11 19:13:37

Overall YJIT is 32.0% faster than interpreted CRuby!
On Railsbench specifically, YJIT is 29.5% faster than CRuby!
The basic "faster" measurement is the geomean of all "headlining" x86 benchmarks on this page.

Headlining Benchmarks

These are "headlining" because the "overall" speedup above is based on these benchmarks specifically.

Speed of each Ruby implementation (iterations/second) relative to the CRuby interpreter. Higher is better.

Other Benchmarks

Speed of each Ruby implementation (iterations/second) relative to the CRuby interpreter. Higher is better.

MicroBenchmarks

Speed of each Ruby implementation (iterations/second) relative to the CRuby interpreter. Higher is better.

Want Raw Graphs and CSV?

Benchmarks Speed Details

Benchmark Memory Usage Details

Number of Iterations and Warmups Tested

bench No JIT warmups No JIT iters MJIT warmups MJIT iters YJIT warmups YJIT iters
activerecord 5 181 75 181 20 181
hexapdf 5 15 20 15
liquid-render 5 156 75 156 20 156
mail 5 123 75 123 20 123
psych-load 5 15 75 15 20 15
railsbench 5 15 75 15 20 15
binarytrees 5 64 75 64 20 64
chunky_png 5 30 75 30 20 30
discourse 5 39 75 39 20 39
erubi 5 47 75 47 20 47
erubi_rails 5 964 75 964 20 964
fannkuchredux 5 15 73 15 20 15
lee 5 22 75 22 20 22
nbody 5 275 75 275 20 275
optcarrot 5 15 75 15 20 15
rubykon 5 15 27 15 20 15
30k_ifelse 5 55 75 55 20 55
30k_methods 5 22 29 22 20 22
cfunc_itself 5 424 75 424 20 424
fib 5 319 75 319 20 319
getivar 5 771 75 771 20 771
keyword_args 5 377 75 377 20 377
respond_to 5 107 75 107 20 107
setivar 5 1781 75 1781 20 1781

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
activerecord (click) 992145 813642 185 1185 0 0% 0 0 0
hexapdf (click) 1099985 923690 600 7897 429 5% 0 0 321
liquid-render (click) 500881 423095 143 1501 92 6% 0 0 5
mail (click) 899089 693574 324 5487 96 1% 0 0 111
psych-load (click) 341841 286772 59 434 1 0% 0 0 0
railsbench (click) 2599761 2105687 1297 9395 224 2% 0 0 26
binarytrees (click) 169489 138404 11 59 0 0% 0 0 0
chunky_png (click) 375313 309129 85 857 0 0% 0 0 0
discourse (click) 6206481 5562397 3102 26625 3119 11% 2 0 140
erubi (click) 329681 277241 10 79 0 0% 0 0 0
erubi_rails (click) 1957393 1597482 262 1896 12 0% 0 0 3
fannkuchredux (click) 183249 149102 8 189 0 0% 0 0 0
lee (click) 378513 315736 45 582 68 11% 0 0 13
nbody (click) 178961 146027 10 158 0 0% 0 0 0
optcarrot (click) 505233 450861 198 3576 20 0% 0 0 0
rubykon (click) 309521 258082 140 1520 2 0% 0 0 0
30k_ifelse (click) 5557265 4358025 9263 57805 0 0% 0 0 0
30k_methods (click) 2173137 1667238 5782 19361 0 0% 0 0 0
cfunc_itself (click) 167185 136659 8 50 0 0% 0 0 0
fib (click) 165777 136267 8 39 0 0% 0 0 0
getivar (click) 167697 139200 8 65 0 0% 0 0 0
keyword_args (click) 169041 138096 9 51 0 0% 0 0 0
respond_to (click) 169041 138206 8 64 0 0% 0 0 0
setivar (click) 167697 137724 8 38 0 0% 0 0 0

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.