YJIT Benchmarks

Details for Benchmarks at 2022-01-04 19:09:55

Overall YJIT is 28.0% faster than interpreted CRuby!
On Railsbench specifically, YJIT is 29.7% 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 5 164 20 164 75 164 20 164
hexapdf 5 15 20 15 20 15
liquid-render 5 136 20 136 75 136 20 136
mail 5 121 20 121 75 121 20 121
psych-load 5 15 20 15 75 15 20 15
railsbench 5 15 20 15 75 15 20 15
binarytrees 5 82 20 82 75 82 20 82
erubi 5 46 20 46 75 46 20 46
erubi_rails 5 547 20 547 75 547 20 547
fannkuchredux 5 15 20 15 52 15 20 15
lee 5 25 20 25 75 25 20 25
nbody 5 339 20 339 75 339 20 339
optcarrot 5 15 20 15 75 15 20 15
rubykon 5 15 20 15 41 15 20 15
30k_ifelse 5 63 20 63 46 63 20 63
30k_methods 5 26 20 26 34 26 20 26
cfunc_itself 5 511 20 511 75 511 20 511
fib 5 382 20 382 75 382 20 382
getivar 5 561 20 561 75 561 20 561
keyword_args 5 428 20 428 75 428 20 428
respond_to 5 131 20 131 75 131 20 131
setivar 5 506 20 506 75 506 20 506

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) 970538 739710 184 1187 0 0% 0 0 0
hexapdf (click) 1118570 912222 599 8120 618 7% 0 0 321
liquid-render (click) 471799 371891 143 1467 95 6% 0 0 1
mail (click) 862250 639020 325 5423 124 2% 0 0 39
psych-load (click) 302570 236324 58 433 1 0% 0 0 0
railsbench (click) 2514978 1900696 1295 9296 266 2% 0 0 26
binarytrees (click) 112677 86078 9 48 0 0% 0 0 0
erubi (click) 272865 213705 8 35 0 0% 0 0 0
erubi_rails (click) 1904098 1421872 260 1979 16 0% 0 0 3
fannkuchredux (click) 128549 97975 6 202 0 0% 0 0 0
lee (click) 331562 260383 44 562 71 12% 0 0 12
nbody (click) 122469 93495 8 148 0 0% 0 0 0
optcarrot (click) 451173 398978 196 3566 20 0% 0 0 0
rubykon (click) 259365 210589 138 1583 2 0% 0 0 0
30k_ifelse (click) 5502181 4306618 9261 57794 0 0% 0 0 0
30k_methods (click) 2117349 1615080 5780 19351 0 0% 0 0 0
cfunc_itself (click) 111653 84923 6 39 0 0% 0 0 0
fib (click) 110117 84531 6 28 0 0% 0 0 0
getivar (click) 111845 87042 6 55 0 0% 0 0 0
keyword_args (click) 112485 85564 7 41 0 0% 0 0 0
respond_to (click) 113253 86048 6 54 0 0% 0 0 0
setivar (click) 111781 85566 6 28 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.