YJIT Benchmarks

Details for Benchmarks at 2022-02-07 19:15:56

Overall YJIT is 32.2% faster than interpreted CRuby!
On Railsbench specifically, YJIT is 30.0% 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 MJIT3.0 warmups MJIT3.0 iters MJIT warmups MJIT iters YJIT warmups YJIT iters
activerecord 5 176 20 176 75 176 20 176
hexapdf 5 15 20 15 20 15
liquid-render 5 155 20 155 75 155 20 155
mail 5 122 20 122 75 122 20 122
psych-load 5 15 20 15 75 15 20 15
railsbench 5 15 20 15 75 15 20 15
binarytrees 5 75 20 75 75 75 20 75
discourse 5 34 20 34 75 34 20 34
erubi 5 47 20 47 75 47 20 47
erubi_rails 5 964 20 964 75 964 20 964
fannkuchredux 5 15 20 15 75 15 20 15
lee 5 21 20 21 75 21 20 21
nbody 5 287 20 287 75 287 20 287
optcarrot 5 15 20 15 75 15 20 15
rubykon 5 15 20 15 38 15 20 15
30k_ifelse 5 55 20 55 75 55 20 55
30k_methods 5 22 20 22 30 22 20 22
cfunc_itself 5 422 20 422 75 422 20 422
fib 5 308 20 308 75 308 20 308
getivar 5 771 20 771 75 771 20 771
keyword_args 5 371 20 371 75 371 20 371
respond_to 5 107 20 107 75 107 20 107
setivar 5 1680 20 1680 75 1680 20 1680

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) 1000145 767174 185 1202 0 0% 0 0 0
hexapdf (click) 1119121 916588 600 8169 613 7% 0 0 321
liquid-render (click) 491153 390718 143 1527 95 6% 0 0 1
mail (click) 900305 674707 324 5502 125 2% 0 0 39
psych-load (click) 337233 267066 59 434 1 0% 0 0 0
railsbench (click) 2537297 1934706 1297 9335 268 2% 0 0 26
binarytrees (click) 151761 116278 11 59 0 0% 0 0 0
discourse (click) 22702225 11403337 3134 285477 261958 91% 2 0 133
erubi (click) 330449 261910 10 79 0 0% 0 0 0
erubi_rails (click) 1910033 1438761 262 1876 18 0% 0 0 3
fannkuchredux (click) 165009 126647 8 189 0 0% 0 0 0
lee (click) 377105 297376 45 587 69 11% 0 0 12
nbody (click) 161041 123479 10 159 0 0% 0 0 0
optcarrot (click) 488273 429109 198 3576 20 0% 0 0 0
rubykon (click) 292689 236352 141 1530 1 0% 0 0 0
30k_ifelse (click) 5540817 4336602 9263 57805 0 0% 0 0 0
30k_methods (click) 2155473 1644735 5782 19362 0 0% 0 0 0
cfunc_itself (click) 149713 114578 8 50 0 0% 0 0 0
fib (click) 148305 114186 8 39 0 0% 0 0 0
getivar (click) 150545 117026 8 66 0 0% 0 0 0
keyword_args (click) 151121 115548 9 52 0 0% 0 0 0
respond_to (click) 151889 116032 8 65 0 0% 0 0 0
setivar (click) 150033 115221 8 39 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.