YJIT Benchmarks

2022-09-21 08:06:32

As of 2022-09-21 08:06:32:

Overall YJIT is 36.7% faster than interpreted CRuby!
On Railsbench specifically, YJIT is 31.7% 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 150.9 0.66% 182.9 4.34% 101.8 0.70% 0.83x 4.38% 1.48x 0.96% 85.99%
hexapdf 3179.7 0.99% 2301.5 1.51% 1.38x 1.80% 82.79%
liquid-render 197.1 0.93% 168.8 1.12% 117.3 1.89% 1.17x 1.45% 1.68x 2.10% 85.97%
mail 172.6 0.14% 174.3 2.44% 153.5 0.17% 0.99x 2.44% 1.12x 0.22% 99.39%
psych-load 2450.4 0.03% 2113.5 0.03% 1910.6 0.03% 1.16x 0.04% 1.28x 0.04% 88.89%
railsbench 2739.6 0.58% 3039.6 0.53% 2080.3 0.92% 0.90x 0.78% 1.32x 1.09% 86.57%
binarytrees 474.9 0.04% 326.0 0.05% 232.3 0.07% 1.46x 0.07% 2.04x 0.08% 100.00%
chunky_png 971.2 0.05% 783.7 0.06% 653.7 0.05% 1.24x 0.08% 1.49x 0.07% 99.95%
erubi 341.0 0.65% 315.1 0.52% 298.0 0.67% 1.08x 0.84% 1.14x 0.93% 100.00%
erubi_rails 26.2 2.85% 27.9 9.44% 18.8 3.63% 0.94x 9.86% 1.39x 4.62% 86.41%
fannkuchredux 5534.3 0.43% 5390.3 0.10% 2290.4 0.07% 1.03x 0.44% 2.42x 0.44% 77.42%
lee 1254.8 0.96% 1022.1 1.64% 944.5 1.27% 1.23x 1.90% 1.33x 1.60% 99.97%
nbody 120.3 0.11% 75.6 0.05% 90.7 0.13% 1.59x 0.12% 1.33x 0.17% 100.00%
optcarrot 6241.3 0.44% 2705.2 0.60% 3473.9 0.62% 2.31x 0.74% 1.80x 0.76% 96.16%
rubykon 12601.5 0.61% 9201.3 0.47% 6841.3 0.67% 1.37x 0.77% 1.84x 0.91% 99.89%
30k_ifelse 2353.7 0.01% 4343.5 3.05% 363.5 0.09% 0.54x 3.05% 6.47x 0.09% 100.00%
30k_methods 6294.9 0.01% 13305.0 3.82% 791.6 0.02% 0.47x 3.82% 7.95x 0.03% 100.00%
cfunc_itself 106.5 0.15% 97.5 0.16% 46.5 0.62% 1.09x 0.22% 2.29x 0.64% 100.00%
fib 239.3 0.04% 79.7 0.03% 62.7 0.04% 3.00x 0.05% 3.82x 0.06% 100.00%
getivar 115.4 0.06% 118.1 0.11% 42.8 0.05% 0.98x 0.13% 2.70x 0.07% 98.73%
keyword_args 293.4 0.08% 227.1 0.10% 54.7 0.15% 1.29x 0.12% 5.36x 0.17% 100.00%
respond_to 274.6 0.23% 325.5 0.21% 30.3 0.44% 0.84x 0.31% 9.06x 0.49% 100.00%
setivar 76.6 0.26% 76.5 0.20% 40.6 0.61% 1.00x 0.33% 1.88x 0.66% 98.35%
str_concat 83.8 0.90% 52.7 6.24% 40.2 1.76% 1.59x 6.31% 2.08x 1.98% 99.96%

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 (2022-09-21T06:48:34Z master ec2d13567e) +YJIT [x86_64-linux]
Tested Ruby version for Ruby latest MJIT: ruby 3.2.0dev (2022-09-21T06:48:34Z master ec2d13567e) +MJIT [x86_64-linux]

Benchmark Memory Usage Details

bench No JIT mem (MiB) MJIT mem (MiB) YJIT mem (MiB)
activerecord 65 72 77
hexapdf 269 418
liquid-render 30 32 37
mail 49 53 57
psych-load 36 45 41
railsbench 102 122 131
binarytrees 30 31 33
chunky_png 45 41 48
erubi 67 101 96
erubi_rails 103 105 122
fannkuchredux 24 26 27
lee 31 41 39
nbody 24 26 27
optcarrot 57 65 65
rubykon 50 65 55
30k_ifelse 68 143 128
30k_methods 58 168 79
cfunc_itself 24 26 27
fib 24 26 27
getivar 25 26 27
keyword_args 24 26 27
respond_to 25 26 27
setivar 24 26 27
str_concat 108 114 116

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 5 130 75 130 20 130
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 67 75 67 20 67
erubi_rails 5 1056 75 1056 20 1056
fannkuchredux 5 15 55 15 20 15
lee 5 21 75 21 20 21
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 67 55 20 55
30k_methods 5 25 22 25 20 25
cfunc_itself 5 433 75 433 20 433
fib 5 313 75 313 20 313
getivar 5 467 75 467 20 467
keyword_args 5 366 75 366 20 366
respond_to 5 656 75 656 20 656
setivar 5 491 75 491 20 491
str_concat 5 491 75 491 20 491

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) 1069623 779168 131 1281 0 0% 0 0 0
hexapdf (click) 1177783 918995 683 8051 10 0% 0 0 0
liquid-render (click) 521271 395285 150 1644 2 0% 0 0 0
mail (click) 932215 650030 374 5408 17 0% 0 0 0
psych-load (click) 356215 261195 68 480 1 0% 0 0 0
railsbench (click) 2732535 1963722 1436 10068 16 0% 0 0 0
binarytrees (click) 177463 128043 13 64 0 0% 0 0 0
chunky_png (click) 398711 288490 88 1003 0 0% 0 0 0
erubi (click) 341111 250594 10 79 0 0% 0 0 0
erubi_rails (click) 2021047 1427752 300 2024 4 0% 0 0 0
fannkuchredux (click) 189815 136544 10 196 0 0% 0 0 0
lee (click) 396407 288628 58 634 0 0% 0 0 0
nbody (click) 186935 135971 12 171 0 0% 0 0 0
optcarrot (click) 531831 441312 207 3592 20 0% 0 0 0
rubykon (click) 329975 255819 144 1583 1 0% 0 0 0
30k_ifelse (click) 5664247 4347294 9265 57804 0 0% 0 0 0
30k_methods (click) 2208503 1655756 5784 19361 0 0% 0 0 0
cfunc_itself (click) 174711 125952 10 49 0 0% 0 0 0
fib (click) 172663 125157 10 38 0 0% 0 0 0
getivar (click) 174519 127294 10 65 0 0% 0 0 0
keyword_args (click) 175031 126190 11 51 0 0% 0 0 0
respond_to (click) 174583 126763 10 64 0 0% 0 0 0
setivar (click) 174391 126064 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.