YJIT Benchmarks

2022-09-19 05:39:24

As of 2022-09-19 05:39:24:

Overall YJIT is 34.4% faster than interpreted CRuby!
On Railsbench specifically, YJIT is 20.8% faster than CRuby!

The basic "faster" measurement is the geomean of all "headlining" benchmarks on this page.

Headlining Benchmarks

These are "headlining" because the "overall" speedup above is based on these benchmarks specifically.

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 No JIT MJIT YJIT activerecord hexapdf liquid-render mail psych-load railsbench Speed of each Ruby implementation (iterations/second) relative to the CRuby interpreter. Higher is better.

MicroBenchmarks

0.0 2.0 4.0 6.0 8.0 No JIT MJIT YJIT 30k_ifelse 30k_methods cfunc_itself fib getivar keyword_args respond_to setivar str_concat Speed of each Ruby implementation (iterations/second) relative to the CRuby interpreter. Higher is better.

Other Benchmarks

0.0 0.5 1.0 1.5 2.0 No JIT MJIT YJIT binarytrees chunky_png erubi erubi_rails fannkuchredux lee nbody optcarrot rubykon Speed of each Ruby implementation (iterations/second) relative to the CRuby interpreter. Higher is better.

Want Raw Graphs and CSV?

Benchmarks Speed Details

The y axis is CRuby's pure-interpreter time for the benchmark divided by the named interpreter's time (YJIT, MJIT, TruffleRuby). Taller bars mean better performance. Each benchmark is scaled to the CRuby's interpreter's mean result, so CRuby's results are always exactly 1.0.

bench No JIT (ms) No JIT RSD MJIT (ms) MJIT RSD YJIT (ms) YJIT RSD MJIT spd MJIT spd RSD YJIT spd YJIT spd RSD % in YJIT
activerecord 149.0 0.81% 177.9 4.11% 103.9 0.60% 0.84x 4.18% 1.43x 1.01% 85.28%
hexapdf 3185.7 1.33% 2347.8 2.27% 1.36x 2.63% 83.00%
liquid-render 196.9 1.61% 169.7 1.13% 116.8 1.87% 1.16x 1.97% 1.69x 2.46% 85.97%
mail 175.9 0.22% 175.7 2.64% 155.4 0.24% 1.00x 2.65% 1.13x 0.33% 99.31%
psych-load 2504.8 0.02% 2141.0 0.05% 1905.0 0.03% 1.17x 0.05% 1.31x 0.04% 88.89%
railsbench 2738.1 0.54% 2957.8 0.86% 2266.3 0.82% 0.93x 1.02% 1.21x 0.98% 86.12%
binarytrees 475.0 0.03% 333.2 0.05% 237.9 0.07% 1.43x 0.06% 2.00x 0.07% 100.00%
chunky_png 975.5 0.08% 794.6 0.12% 655.2 0.05% 1.23x 0.14% 1.49x 0.10% 99.95%
erubi 336.3 0.62% 316.7 2.77% 297.1 0.72% 1.06x 2.84% 1.13x 0.95% 100.00%
erubi_rails 26.4 2.57% 29.5 10.06% 19.1 3.51% 0.89x 10.38% 1.38x 4.35% 86.35%
fannkuchredux 5299.4 0.16% 5343.4 0.04% 2366.5 0.14% 0.99x 0.17% 2.24x 0.21% 77.42%
lee 1267.6 0.91% 1031.5 1.51% 961.3 1.15% 1.23x 1.76% 1.32x 1.47% 99.97%
nbody 120.4 0.42% 75.6 0.05% 91.7 0.11% 1.59x 0.42% 1.31x 0.43% 100.00%
optcarrot 6097.5 0.43% 2700.6 0.63% 3530.1 0.60% 2.26x 0.76% 1.73x 0.74% 96.16%
rubykon 12659.5 0.66% 9580.7 0.37% 6856.4 0.63% 1.32x 0.76% 1.85x 0.91% 99.61%
30k_ifelse 2286.1 0.01% 4389.2 2.75% 359.2 0.10% 0.52x 2.75% 6.36x 0.10% 100.00%
30k_methods 6247.3 0.01% 12853.8 3.85% 783.9 0.02% 0.49x 3.85% 7.97x 0.02% 100.00%
cfunc_itself 107.7 0.17% 98.6 0.12% 46.6 0.64% 1.09x 0.21% 2.31x 0.66% 100.00%
fib 239.9 0.10% 79.2 0.03% 63.7 0.06% 3.03x 0.11% 3.76x 0.12% 100.00%
getivar 102.9 0.12% 104.4 0.13% 42.8 0.05% 0.99x 0.18% 2.41x 0.13% 98.73%
keyword_args 291.5 0.08% 226.7 0.05% 53.8 0.15% 1.29x 0.10% 5.42x 0.17% 100.00%
respond_to 277.5 0.15% 259.8 0.21% 29.9 0.43% 1.07x 0.26% 9.27x 0.45% 100.00%
setivar 77.9 0.17% 77.8 0.18% 49.9 0.28% 1.00x 0.25% 1.56x 0.33% 98.01%
str_concat 83.4 0.85% 54.5 7.91% 47.3 1.48% 1.53x 7.95% 1.77x 1.70% 99.97%

RSD is relative standard deviation - the standard deviation divided by the mean, expressed as a percentage.
% in YJIT is the percentage of instructions that complete in YJIT rather than exiting to the non-JITted interpreter. YJIT performs better when this is higher.
Speedup is relative to interpreted CRuby. So an "MJIT speedup" of 1.21x means MJIT runs at 1.21 times the iters/second of CRuby with JIT disabled.

You can find our benchmark code in the yjit-bench Github repo and the yjit-extra-benchmarks Github repo.
Our benchmark-runner and reporting code is in the yjit-metrics Github repo.

Tested Ruby version for YJIT and No-JIT: ruby 3.2.0dev (2022-09-19T05:15:13Z master 12e5e5b573) +YJIT [x86_64-linux]
Tested Ruby version for Ruby latest MJIT: ruby 3.2.0dev (2022-09-19T05:15:13Z master 12e5e5b573) +MJIT [x86_64-linux]

Benchmark Memory Usage Details

bench No JIT mem (MiB) MJIT mem (MiB) YJIT mem (MiB)
activerecord 65 72 76
hexapdf 271 396
liquid-render 30 33 37
mail 49 54 56
psych-load 35 44 41
railsbench 102 123 131
binarytrees 30 31 33
chunky_png 44 42 49
erubi 74 79 95
erubi_rails 101 105 122
fannkuchredux 24 26 27
lee 33 41 39
nbody 24 26 27
optcarrot 57 65 65
rubykon 51 63 55
30k_ifelse 67 145 128
30k_methods 58 166 79
cfunc_itself 25 26 27
fib 24 26 27
getivar 25 25 27
keyword_args 24 25 27
respond_to 25 26 27
setivar 25 26 27
str_concat 98 114 101

Memory is shown in mebibytes (1024 * 1024 bytes.)

Older YJIT allocated an additional 256MiB for generated code. Current YJIT allocates executable memory on demand, so this overhead should no longer be present.

Number of Iterations and Warmups Tested

bench No JIT warmups No JIT iters MJIT warmups MJIT iters YJIT warmups YJIT iters
activerecord 5 192 75 192 20 192
hexapdf 5 15 20 15
liquid-render 5 171 75 171 20 171
mail 5 128 75 128 20 128
psych-load 5 15 75 15 20 15
railsbench 5 15 75 15 20 15
binarytrees 5 84 75 84 20 84
chunky_png 5 30 75 30 20 30
erubi 5 68 75 68 20 68
erubi_rails 5 1057 75 1057 20 1057
fannkuchredux 5 15 56 15 20 15
lee 5 20 75 20 20 20
nbody 5 264 75 264 20 264
optcarrot 5 15 75 15 20 15
rubykon 5 15 32 15 20 15
30k_ifelse 5 55 70 55 20 55
30k_methods 5 25 23 25 20 25
cfunc_itself 5 429 75 429 20 429
fib 5 318 75 318 20 318
getivar 5 467 75 467 20 467
keyword_args 5 371 75 371 20 371
respond_to 5 663 75 663 20 663
setivar 5 402 75 402 20 402
str_concat 5 421 75 421 20 421

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 column headings for descriptions of each statistic.

bench Exit Report Inline Outlined Comp iSeqs Comp Blocks Inval Inval Ratio Bind Alloc Bind Set Const Bumps
activerecord (click) 1065847 776445 131 1270 0 0% 0 0 0
hexapdf (click) 1176247 917757 682 8039 9 0% 0 0 0
liquid-render (click) 529079 401144 150 1725 2 0% 0 0 0
mail (click) 924279 644814 374 5355 17 0% 0 0 0
psych-load (click) 355447 260647 68 474 1 0% 0 0 0
railsbench (click) 2720503 1955090 1436 10014 16 0% 0 0 0
binarytrees (click) 177079 127640 13 64 0 0% 0 0 0
chunky_png (click) 396791 287262 88 997 0 0% 0 0 0
erubi (click) 341111 250519 10 79 0 0% 0 0 0
erubi_rails (click) 2013175 1422243 300 2028 4 0% 0 0 0
fannkuchredux (click) 189879 136594 10 196 0 0% 0 0 0
lee (click) 395511 287802 58 634 0 0% 0 0 0
nbody (click) 187319 136374 12 171 0 0% 0 0 0
optcarrot (click) 531831 441312 207 3592 20 0% 0 0 0
rubykon (click) 328759 254726 144 1568 1 0% 0 0 0
30k_ifelse (click) 5664183 4347244 9265 57804 0 0% 0 0 0
30k_methods (click) 2208503 1655756 5784 19361 0 0% 0 0 0
cfunc_itself (click) 174391 125599 10 49 0 0% 0 0 0
fib (click) 173047 125560 10 38 0 0% 0 0 0
getivar (click) 174519 127294 10 65 0 0% 0 0 0
keyword_args (click) 175415 126593 11 51 0 0% 0 0 0
respond_to (click) 174647 126813 10 64 0 0% 0 0 0
setivar (click) 174455 126114 10 38 0 0% 0 0 0
str_concat (click) 174391 126446 12 62 0 0% 0 0 0

YJIT stats correspond to the YJIT stats exit report.

Raw JSON data files

All graphs and table data in this page comes from processing these data files, which come from benchmark runs.