YJIT Benchmarks

2022-07-31 05:14:21

As of 2022-07-31 05:14:21:

Overall YJIT is 33.6% faster than interpreted CRuby!
On Railsbench specifically, YJIT is 24.4% 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 No JIT MJIT3.1 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 1.0 2.0 3.0 4.0 5.0 6.0 7.0 No JIT MJIT3.1 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 MJIT3.1 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 MJIT3.1 (ms) MJIT3.1 RSD YJIT (ms) YJIT RSD MJIT3.1 spd MJIT3.1 spd RSD YJIT spd YJIT spd RSD % in YJIT
activerecord 149.6 0.36% 152.6 0.98% 103.3 0.61% 0.98x 1.04% 1.45x 0.71% 85.94%
hexapdf 3110.5 0.92% 2298.3 3.01% 1.35x 3.15% 83.65%
liquid-render 197.1 0.98% 170.3 1.63% 124.9 1.73% 1.16x 1.90% 1.58x 1.99% 85.90%
mail 178.9 0.14% 174.6 1.96% 157.8 0.13% 1.02x 1.97% 1.13x 0.19% 99.36%
psych-load 2506.7 0.03% 2210.6 0.02% 1922.4 0.03% 1.13x 0.04% 1.30x 0.05% 88.89%
railsbench 2813.5 0.53% 2831.5 0.50% 2261.9 0.76% 0.99x 0.73% 1.24x 0.93% 86.54%
binarytrees 479.1 0.04% 312.6 0.31% 369.0 0.05% 1.53x 0.31% 1.30x 0.06% 84.27%
chunky_png 956.0 0.03% 766.1 0.03% 638.0 0.12% 1.25x 0.04% 1.50x 0.12% 99.95%
erubi 376.2 0.65% 339.6 0.54% 316.6 0.61% 1.11x 0.84% 1.19x 0.89% 100.00%
erubi_rails 26.9 2.73% 27.0 2.73% 19.3 3.51% 1.00x 3.86% 1.39x 4.45% 86.40%
fannkuchredux 6939.8 0.10% 4046.3 0.04% 6947.0 0.09% 1.72x 0.11% 1.00x 0.14% 0.02%
lee 1294.8 0.79% 1033.0 1.39% 938.9 0.58% 1.25x 1.60% 1.38x 0.98% 99.97%
nbody 125.9 0.11% 76.3 0.24% 93.5 0.09% 1.65x 0.26% 1.35x 0.14% 100.00%
optcarrot 6293.4 0.42% 2802.5 0.63% 3420.7 0.61% 2.25x 0.76% 1.84x 0.74% 96.16%
rubykon 12679.6 0.55% 8096.7 0.49% 6756.9 0.55% 1.57x 0.73% 1.88x 0.77% 99.67%
30k_ifelse 2241.6 0.02% 3555.5 4.03% 353.4 0.09% 0.63x 4.03% 6.34x 0.10% 100.00%
30k_methods 6715.8 0.06% 9917.7 18.35% 873.6 0.02% 0.68x 18.35% 7.69x 0.07% 100.00%
cfunc_itself 107.5 0.12% 69.6 0.86% 44.7 0.66% 1.54x 0.87% 2.40x 0.67% 100.00%
fib 243.5 0.02% 86.3 0.30% 63.3 0.04% 2.82x 0.30% 3.85x 0.05% 100.00%
getivar 116.9 0.09% 32.8 59.34% 44.3 0.12% 3.56x 59.34% 2.64x 0.15% 99.02%
keyword_args 287.6 0.09% 224.4 0.65% 52.7 0.35% 1.28x 0.66% 5.46x 0.36% 100.00%
respond_to 276.1 0.12% 233.6 0.67% 193.8 0.09% 1.18x 0.68% 1.42x 0.15% 100.00%
setivar 77.4 0.14% 20.9 63.58% 46.4 0.12% 3.71x 63.58% 1.67x 0.18% 99.14%
str_concat 85.2 0.93% 50.6 1.55% 43.4 1.69% 1.68x 1.81% 1.96x 1.93% 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 it 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.
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 (2022-07-30T18:18:59Z master 118368c1dd) +YJIT [x86_64-linux]
Tested Ruby version for Ruby 3.1 MJIT: ruby 3.2.0dev (2022-07-30T18:18:59Z master 118368c1dd) +MJIT [x86_64-linux]
(We got much better MJIT results with released than prerelease, so we used those.)

Benchmark Memory Usage Details

bench No JIT mem (MiB) MJIT3.1 mem (MiB) YJIT mem (MiB)
activerecord 65 66 76
hexapdf 346 414
liquid-render 31 35 37
mail 49 53 56
psych-load 36 45 41
railsbench 104 107 133
binarytrees 31 33 33
chunky_png 44 45 48
erubi 78 121 89
erubi_rails 102 103 122
fannkuchredux 25 27 27
lee 33 40 39
nbody 24 26 27
optcarrot 58 66 65
rubykon 53 58 57
30k_ifelse 68 170 127
30k_methods 57 176 76
cfunc_itself 25 27 27
fib 24 27 27
getivar 25 27 27
keyword_args 25 27 27
respond_to 25 27 27
setivar 25 26 27
str_concat 122 111 117

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

By default, YJIT allocates an additional 256MiB for generated code. The additional size allocated can be tuned with a command-line parameter, but allocating too little memory will result in YJIT no longer JITting new functions, so performance reverts toward interpreter speeds.

Number of Iterations and Warmups Tested

bench No JIT warmups No JIT iters MJIT3.1 warmups MJIT3.1 iters YJIT warmups YJIT iters
activerecord 5 190 75 190 20 190
hexapdf 5 15 20 15
liquid-render 5 157 75 157 20 157
mail 5 128 75 128 20 128
psych-load 5 15 75 15 20 15
railsbench 5 15 75 15 20 15
binarytrees 5 64 75 64 20 64
chunky_png 5 30 75 30 20 30
erubi 5 62 75 62 20 62
erubi_rails 5 1007 75 1007 20 1007
fannkuchredux 5 15 74 15 20 15
lee 5 21 75 21 20 21
nbody 5 262 75 262 20 262
optcarrot 5 15 75 15 20 15
rubykon 5 15 35 15 20 15
30k_ifelse 5 56 75 56 20 56
30k_methods 5 22 30 22 20 22
cfunc_itself 5 446 75 446 20 446
fib 5 315 75 315 20 315
getivar 5 611 75 611 20 611
keyword_args 5 377 75 377 20 377
respond_to 5 104 75 104 20 104
setivar 5 962 75 962 20 962
str_concat 5 467 75 467 20 467

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) 1035611 779646 131 1264 0 0% 0 0 0
hexapdf (click) 1150747 925327 682 8014 9 0% 0 0 0
liquid-render (click) 518491 403971 150 1671 2 0% 0 0 0
mail (click) 922523 665614 374 5549 17 0% 0 0 0
psych-load (click) 348059 263768 68 482 1 0% 0 0 0
railsbench (click) 2654363 1967044 1436 10009 16 0% 0 0 0
binarytrees (click) 172187 128374 13 58 0 0% 0 0 0
chunky_png (click) 386971 289411 88 998 0 0% 0 0 0
erubi (click) 323547 246809 10 79 0 0% 0 0 0
erubi_rails (click) 1962779 1430531 300 2026 4 0% 0 0 0
fannkuchredux (click) 186011 138504 10 188 0 0% 0 0 0
lee (click) 385819 291379 58 632 0 0% 0 0 0
nbody (click) 182811 137569 12 171 0 0% 0 0 0
optcarrot (click) 514203 442736 207 3588 20 0% 0 0 0
rubykon (click) 321307 256707 144 1573 1 0% 0 0 0
30k_ifelse (click) 5561243 4348842 9265 57804 0 0% 0 0 0
30k_methods (click) 2176411 1657304 5784 19361 0 0% 0 0 0
cfunc_itself (click) 170715 127147 10 49 0 0% 0 0 0
fib (click) 169179 126755 10 38 0 0% 0 0 0
getivar (click) 170907 128842 10 65 0 0% 0 0 0
keyword_args (click) 171547 127788 11 51 0 0% 0 0 0
respond_to (click) 172379 128305 10 64 0 0% 0 0 0
setivar (click) 170843 127662 10 38 0 0% 0 0 0
str_concat (click) 170843 128069 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.