YJIT Benchmarks

Details for Benchmarks at 2021-10-18 12:11:21

Overall YJIT is 23.2% faster than interpreted CRuby!
On Railsbench specifically, YJIT is 21.2% 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 YJIT warmups YJIT iters
activerecord 10 128 10 123 10 177
hexapdf 10 10 10 10 10 10
jekyll 10 10 10 10 10 10
liquid-render 10 106 10 112 10 145
mail 10 119 10 123 10 133
psych-load 10 10 10 10 10 10
railsbench 10 10 10 10 10 10
binarytrees 10 45 10 69 10 60
fannkuchredux 10 10 10 10 10 10
lee 10 10 10 13 10 15
nbody 10 174 10 331 10 256
optcarrot 10 10 10 10 10 10
30k_ifelse 10 10 10 10 10 51
30k_methods 10 10 10 10 10 17
cfunc_itself 10 204 10 322 10 495
fib 10 94 10 307 10 375
getivar 10 224 10 211 10 530
keyword_args 10 77 10 97 10 416
respond_to 10 76 10 88 10 109
setivar 10 284 10 259 10 458

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) 1113178 866931 185 1156 0 0% 0 0 0
hexapdf (click) 1066164 892200 604 7832 619 7% 0 0 320
jekyll (click) 1960922 1631125 336 3708 173 4% 0 0 0
liquid-render (click) 459220 375674 143 1296 94 7% 0 0 1
mail (click) 977306 755328 337 4997 129 2% 0 0 39
psych-load (click) 185066 149179 56 409 1 0% 0 0 0
railsbench (click) 2562993 1997530 1293 9557 263 2% 0 0 26
binarytrees (click) 113853 90204 9 49 0 0% 0 0 0
fannkuchredux (click) 129341 103435 6 203 0 0% 0 0 0
lee (click) 235178 181797 85 1091 114 10% 1 0 13
nbody (click) 123005 97556 8 149 0 0% 0 0 0
optcarrot (click) 458877 429332 196 3567 20 0% 0 0 0
30k_ifelse (click) 5560701 4890509 9261 57795 0 0% 0 0 0
30k_methods (click) 2117693 1816572 5780 19352 0 0% 0 0 0
cfunc_itself (click) 111549 88019 6 40 0 0% 0 0 0
fib (click) 110781 88216 6 29 0 0% 0 0 0
getivar (click) 112189 90811 6 56 0 0% 0 0 0
keyword_args (click) 112253 88627 7 42 0 0% 0 0 0
respond_to (click) 113853 89563 6 55 0 0% 0 0 0
setivar (click) 112189 89120 6 29 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.


Far Too Many Details

These are links to HTML reports on GitHub with far, far too much detail about the benchmarking runs per-platform.


Autogenerated by continuous_reporting/generate_and_upload_reports.rb script.