YJIT Benchmarks

Details for Benchmarks at 2021-12-25 07:12:08

Overall YJIT is 25.9% faster than interpreted CRuby!
On Railsbench specifically, YJIT is 26.7% 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 5 188 20 188 75 188 20 188
hexapdf 5 15 20 15 20 15
liquid-render 5 160 20 160 75 160 20 160
mail 5 142 20 142 75 142 20 142
psych-load 5 15 20 15 75 15 20 15
railsbench 5 15 20 15 75 15 20 15
binarytrees 5 88 20 88 75 88 20 88
erubi 5 52 20 52 75 52 20 52
erubi_rails 5 596 20 596 75 596 20 596
fannkuchredux 5 15 20 15 54 15 20 15
jekyll 5 15 20 15 36 15 20 15
lee 5 26 20 26 75 26 20 26
nbody 5 341 20 341 75 341 20 341
optcarrot 5 15 20 15 75 15 20 15
rubykon 5 15 20 15 43 15 20 15
30k_ifelse 5 64 20 64 70 64 20 64
30k_methods 5 26 20 26 44 26 20 26
cfunc_itself 5 506 20 506 75 506 20 506
fib 5 384 20 384 75 384 20 384
getivar 5 564 20 564 75 564 20 564
keyword_args 5 423 20 423 75 423 20 423
respond_to 5 128 20 128 75 128 20 128
setivar 5 503 20 503 75 503 20 503

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) 978410 746248 185 1194 0 0% 0 0 0
hexapdf (click) 1122218 916846 598 8107 620 7% 0 0 321
liquid-render (click) 474871 375054 143 1461 96 6% 0 0 1
mail (click) 870954 646705 325 5440 124 2% 0 0 39
psych-load (click) 299041 235617 57 423 1 0% 0 0 0
railsbench (click) 2517858 1904070 1295 9290 266 2% 0 0 26
binarytrees (click) 112933 86584 9 48 0 0% 0 0 0
erubi (click) 279009 219396 8 35 0 0% 0 0 0
erubi_rails (click) 2461986 1774494 260 1968 17 0% 0 0 3
fannkuchredux (click) 129061 98707 6 202 0 0% 0 0 0
jekyll (click) 1845802 1483315 338 3730 185 4% 0 0 0
lee (click) 336993 265350 43 557 70 12% 0 0 12
nbody (click) 122469 93898 8 148 0 0% 0 0 0
optcarrot (click) 451685 399710 196 3566 20 0% 0 0 0
rubykon (click) 254693 207061 138 1528 2 0% 0 0 0
30k_ifelse (click) 5502053 4306908 9261 57794 0 0% 0 0 0
30k_methods (click) 2117221 1615370 5780 19351 0 0% 0 0 0
cfunc_itself (click) 112037 85542 6 39 0 0% 0 0 0
fib (click) 110117 84934 6 28 0 0% 0 0 0
getivar (click) 111845 87445 6 55 0 0% 0 0 0
keyword_args (click) 112485 85967 7 41 0 0% 0 0 0
respond_to (click) 113637 86667 6 54 0 0% 0 0 0
setivar (click) 112165 86185 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.