YJIT Benchmarks

Details for Benchmarks at 2022-10-07 20:10:39

Overall YJIT is 36.1% faster than interpreted CRuby!
On Railsbench specifically, YJIT is 29.2% 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 YJIT warmups YJIT iters
activerecord 5 192 5 192
hexapdf 5 15 5 15
liquid-render 5 175 5 175
mail 5 129 5 129
psych-load 5 15 5 15
railsbench 5 15 5 15
binarytrees 5 84 5 84
chunky_png 5 31 5 31
erubi 5 67 5 67
erubi_rails 5 1014 5 1014
fannkuchredux 5 15 5 15
lee 5 20 5 20
nbody 5 217 5 217
optcarrot 5 15 5 15
rubykon 5 15 5 15
30k_ifelse 5 57 5 57
30k_methods 5 24 5 24
cfunc_itself 5 456 5 456
fib 5 324 5 324
getivar 5 467 5 467
keyword_args 5 372 5 372
respond_to 5 668 5 668
setivar 5 420 5 420
str_concat 5 469 5 469

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) 2480337 1542623 23 265 6 2% 0 0 0
hexapdf (click) 1887379 1384673 872 9352 181 1% 0 0 0
liquid-render (click) 742547 513353 154 1659 34 2% 0 0 0
mail (click) 1479763 992577 385 4901 174 3% 0 0 0
psych-load (click) 938131 627241 68 435 18 4% 0 0 0
railsbench (click) 5395025 3447690 1784 12465 470 3% 0 0 0
binarytrees (click) 318163 203226 14 70 4 5% 0 0 0
chunky_png (click) 959571 646987 87 1011 31 3% 0 0 0
erubi (click) 891091 600897 11 76 4 5% 0 0 0
erubi_rails (click) 4478545 2771042 327 2428 104 4% 0 0 0
fannkuchredux (click) 330771 212632 11 200 4 2% 0 0 0
lee (click) 974419 659805 80 744 25 3% 0 0 0
nbody (click) 329875 213344 13 179 6 3% 0 0 0
optcarrot (click) 752979 591725 201 3773 93 2% 0 0 0
rubykon (click) 474963 333885 143 1568 32 2% 0 0 0
30k_ifelse (click) 5624979 4414714 9266 57127 4 0% 0 0 0
30k_methods (click) 2271827 1727826 5785 19356 4 0% 0 0 0
cfunc_itself (click) 315603 201078 11 53 4 7% 0 0 0
fib (click) 309971 198133 11 42 4 9% 0 0 0
getivar (click) 316947 203612 11 69 4 5% 0 0 0
keyword_args (click) 316371 201719 12 55 4 7% 0 0 0
respond_to (click) 314963 201513 11 55 5 9% 0 0 0
setivar (click) 312211 199275 11 42 4 9% 0 0 0
str_concat (click) 315859 201927 13 68 6 8% 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.