YJIT Benchmarks

Details for Benchmarks at 2024-05-26 06:08:15 GMT

YJIT metrics from the yjit-bench suite using Ruby d2c4363e0c.

¶Performance on Headline Benchmarks

Select Platform

Xeon x86_64 AWS Graviton ARM64

X

Speed of each Ruby implementation relative to the baseline CRuby measurement. Higher is better.

¶Memory Usage on Headline Benchmarks

Select Platform

Xeon x86_64 AWS Graviton ARM64

X

Memory usage of each Ruby implementation relative to the baseline CRuby measurement. Lower is better.

¶Performance on Other Benchmarks

Select Platform

Xeon x86_64 AWS Graviton ARM64

X

Speed of each Ruby implementation relative to the baseline CRuby measurement. Higher is better.

¶Memory Usage on Other Benchmarks

Select Platform

Xeon x86_64 AWS Graviton ARM64

X

Memory usage of each Ruby implementation relative to the baseline CRuby measurement. Lower is better.

¶Performance on MicroBenchmarks

Select Platform

Xeon x86_64 AWS Graviton ARM64

X

Speed of each Ruby implementation relative to the baseline CRuby measurement. Higher is better.

¶Memory Usage on MicroBenchmarks

Select Platform

Xeon x86_64 AWS Graviton ARM64

X

Memory usage of each Ruby implementation relative to the baseline CRuby measurement. Lower is better.

¶Want Raw Graphs and CSV?

• Headlining benchmarks (x86_64)
• All benchmarks (x86_64)
• CSV speed results (x86_64)
• Headlining benchmarks (aarch64)
• All benchmarks (aarch64)
• CSV speed results (aarch64)

¶Benchmarks Speed Details

Select Platform

Xeon x86_64 AWS Graviton ARM64

X

¶Benchmark Memory Usage Details

Select Platform

Xeon x86_64 AWS Graviton ARM64

X

bench	CRuby 3.3.1 mem (MiB)	CRuby 3.4.0dev mem (MiB)	YJIT 3.3.1 mem (MiB)	YJIT 3.4.0dev mem (MiB)	Inline Code	Outlined Code	YJIT Mem overhead
activerecord	62	68	79	75	2	2	10.5%
chunky-png	40	50	47	48	1	1	-4.4%
erubi-rails	104	109	104	111	2	2	1.9%
hexapdf	156	145	154	270	2	2	85.3%
liquid-c	35	37	40	41	1	1	11.4%
liquid-compile	33	35	36	37	1	1	7.7%
liquid-render	36	37	41	42	1	1	15.3%
lobsters	276	275	315	324	8	6	17.8%
mail	50	68	54	71	1	1	4.7%
psych-load	33	34	34	36	1	1	6.3%
railsbench	101	107	115	115	3	3	7.3%
rubocop	89	92	120	124	6	5	35.0%
ruby-lsp	83	72	88	75	1	1	3.2%
sequel	36	39	40	42	1	1	7.3%
binarytrees	25	26	24	27	1	1	4.1%
blurhash	21	22	22	23	1	1	1.5%
erubi	29	31	31	32	1	1	4.7%
etanni	23	25	23	25	1	1	-1.2%
fannkuchredux	20	21	20	21	1	1	0.8%
fluentd	606	609	608	660	1	1	8.4%
graphql	39	41	42	44	1	1	7.5%
graphql-native	37	40	40	42	1	1	6.7%
lee	34	34	36	35	1	1	3.0%
matmul	32	26	32	26	1	1	0.6%
nbody	20	21	20	21	1	1	0.7%
nqueens	20	22	20	22	1	1	3.3%
optcarrot	59	68	60	76	1	1	12.4%
protoboeuf	29	33	29	34	1	1	2.3%
rack	30	32	32	34	1	1	5.6%
ruby-json	21	22	21	22	1	1	1.0%
rubykon	48	47	51	47	1	1	-0.3%
sudoku	20	21	20	21	1	1	1.0%
tinygql	27	29	29	31	1	1	7.4%
30k_ifelse	60	146	90	154	5	5	5.8%
30k_methods	49	92	58	96	2	2	3.8%
cfunc_itself	20	21	20	21	1	1	0.6%
fib	20	21	20	21	1	1	0.9%
getivar	20	21	20	21	1	1	0.9%
keyword_args	20	21	20	21	1	1	0.9%
respond_to	20	21	20	21	1	1	1.0%
ruby-xor	21	22	21	22	1	1	0.6%
setivar	20	21	20	21	1	1	0.6%
setivar_object	20	21	20	21	1	1	0.6%
setivar_young	20	21	20	21	1	1	0.6%
str_concat	46	47	46	47	1	1	0.6%
throw	20	21	20	21	1	1	0.6%

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

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.

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.

• aarch64_prev_ruby_no_jit
• aarch64_prev_ruby_yjit
• aarch64_prod_ruby_no_jit
• aarch64_prod_ruby_with_yjit
• aarch64_yjit_stats
• x86_64_prev_ruby_no_jit
• x86_64_prev_ruby_yjit
• x86_64_prod_ruby_no_jit
• x86_64_prod_ruby_with_yjit
• x86_64_yjit_stats