[WIP] Try applying prefix perfect hashing to long strings

[jviotti]

Signed-off-by: Juan Cruz Viotti [email protected]

[jviotti]

@michaelmior Can you trigger a full benchmark run on this branch so we see how the results look like?

[github-actions]

Benchmark (linux/llvm)

Benchmark suite	Current: eedfb8e	Previous: 948ff42	Ratio
Compiler_Draft6_AdaptiveCard	68596540133.000015 ns/iter	68312483298.999954 ns/iter	1.00
Compiler_2019_09_OMC_JSON_V2	11096021441.999994 ns/iter	11027577914.999994 ns/iter	1.01
Evaluator_Draft4_Meta_1_No_Callback	287.02182420632744 ns/iter	283.29378537909406 ns/iter	1.01
Evaluator_Draft4_Required_Properties	674.3950864224112 ns/iter	632.4423405808556 ns/iter	1.07
Evaluator_Draft4_Many_Optional_Properties_Minimal_Match	32.208382546147114 ns/iter	33.192458142463074 ns/iter	0.97
Evaluator_Draft4_Few_Optional_Properties_Minimal_Match	14.006250490433253 ns/iter	13.700405805883841 ns/iter	1.02
Evaluator_Draft4_Items_Schema	586.7326471202888 ns/iter	602.5183530511814 ns/iter	0.97
Evaluator_Draft4_Nested_Object	5.608376677477891 ns/iter	5.29143462267445 ns/iter	1.06
Evaluator_Draft4_Properties_Triad_Optional	903.0460415765921 ns/iter	909.7315316208591 ns/iter	0.99
Evaluator_Draft4_Properties_Triad_Closed	724.3363530532328 ns/iter	704.8883847375886 ns/iter	1.03
Evaluator_Draft4_Properties_Triad_Required	901.690103340391 ns/iter	870.563567254097 ns/iter	1.04
Evaluator_Draft4_Properties_Closed	83.25563971666479 ns/iter	84.72792529301914 ns/iter	0.98
Evaluator_Draft4_Non_Recursive_Ref	29.246666960170103 ns/iter	32.42864832421251 ns/iter	0.90
Evaluator_Draft4_Pattern_Properties_True	241.85851901583533 ns/iter	238.17980650851408 ns/iter	1.02
Evaluator_Draft4_Ref_To_Single_Property	13.993164473480048 ns/iter	13.386196486430197 ns/iter	1.05
Evaluator_Draft4_Additional_Properties_Type	53.48384929094504 ns/iter	61.42778935417907 ns/iter	0.87
Evaluator_Draft4_Nested_Oneof	89.23402448405889 ns/iter	87.09661132245628 ns/iter	1.02
Evaluator_Draft4_Short_Enum	17.737105804614668 ns/iter	16.541326548371554 ns/iter	1.07
Evaluator_Draft4_Long_Enum	19.65938255813928 ns/iter	20.60598578090524 ns/iter	0.95
Evaluator_Draft4_Long_Enum_Short_Strings	17.804522591311468 ns/iter	16.527633495163105 ns/iter	1.08
Evaluator_Draft4_Type_Object	10.265734120365417 ns/iter	10.675890973243625 ns/iter	0.96
Evaluator_Draft6_Property_Names	159.38770500269715 ns/iter	135.21534181685047 ns/iter	1.18
Evaluator_Draft7_If_Then_Else	41.04596502633076 ns/iter	40.264602685940325 ns/iter	1.02
Evaluator_Draft7_Vercel_1	46888.314678074166 ns/iter	46109.40225884788 ns/iter	1.02
Evaluator_Draft7_Helm_Chart_Lock_1	281.121773365759 ns/iter	290.27327204713333 ns/iter	0.97
Evaluator_Draft7_Helm_Chart_Lock_2	274.48599004714737 ns/iter	285.5089956124682 ns/iter	0.96
Evaluator_Draft7_Jasmine_1	186.60235312017082 ns/iter	183.02722014496152 ns/iter	1.02
Evaluator_Draft7_Dependabot_1	1213.752554345237 ns/iter	1213.7633163001685 ns/iter	1.00
Evaluator_2019_09_Unevaluated_Properties	52.6485211349994 ns/iter	53.5883776303539 ns/iter	0.98
Evaluator_2019_09_OMC_JSON_V2_1	1468.248543393476 ns/iter	1453.4585640775135 ns/iter	1.01
Evaluator_2020_12_Dynamic_Ref	982.2173287675413 ns/iter	555.3676392593713 ns/iter	1.77
Evaluator_2020_12_Dynamic_Ref_Single	10.596666268324332 ns/iter	10.29100527968794 ns/iter	1.03
Evaluator_2020_12_CQL_1	4640.775278137335 ns/iter	3675.2992595013147 ns/iter	1.26
Evaluator_2020_12_OpenAPI	200629.27231253422 ns/iter	198604.35611510076 ns/iter	1.01

This comment was automatically generated by workflow using github-action-benchmark.

[github-actions]

Benchmark (macos/llvm)

Benchmark suite	Current: eedfb8e	Previous: 948ff42	Ratio
Compiler_Draft6_AdaptiveCard	52587980500.000015 ns/iter	61551956292 ns/iter	0.85
Compiler_2019_09_OMC_JSON_V2	8766306999.999983 ns/iter	10721463792.00001 ns/iter	0.82
Evaluator_Draft4_Meta_1_No_Callback	178.3982751727045 ns/iter	218.58423096866224 ns/iter	0.82
Evaluator_Draft4_Required_Properties	408.232776378486 ns/iter	518.066517757114 ns/iter	0.79
Evaluator_Draft4_Many_Optional_Properties_Minimal_Match	19.394285565245443 ns/iter	29.246656981953603 ns/iter	0.66
Evaluator_Draft4_Few_Optional_Properties_Minimal_Match	7.900929154494332 ns/iter	10.36342343954891 ns/iter	0.76
Evaluator_Draft4_Items_Schema	271.3148910970938 ns/iter	397.875918161459 ns/iter	0.68
Evaluator_Draft4_Nested_Object	3.2956612047700413 ns/iter	5.0168518991750854 ns/iter	0.66
Evaluator_Draft4_Properties_Triad_Optional	795.6940817821494 ns/iter	1193.4733223289068 ns/iter	0.67
Evaluator_Draft4_Properties_Triad_Closed	627.8121637179256 ns/iter	984.0590789701389 ns/iter	0.64
Evaluator_Draft4_Properties_Triad_Required	784.2035776243488 ns/iter	1291.783498677157 ns/iter	0.61
Evaluator_Draft4_Properties_Closed	46.52591351788647 ns/iter	71.25083602950475 ns/iter	0.65
Evaluator_Draft4_Non_Recursive_Ref	10.127182502402487 ns/iter	14.247365219812387 ns/iter	0.71
Evaluator_Draft4_Pattern_Properties_True	167.75244241489358 ns/iter	259.2083879612843 ns/iter	0.65
Evaluator_Draft4_Ref_To_Single_Property	7.424063405605765 ns/iter	10.786996372168245 ns/iter	0.69
Evaluator_Draft4_Additional_Properties_Type	17.281543375811363 ns/iter	23.766955624504845 ns/iter	0.73
Evaluator_Draft4_Nested_Oneof	55.58319185997478 ns/iter	79.32997939620354 ns/iter	0.70
Evaluator_Draft4_Short_Enum	9.615669694169696 ns/iter	11.735057082014663 ns/iter	0.82
Evaluator_Draft4_Long_Enum	31.607759110437282 ns/iter	36.63894305072789 ns/iter	0.86
Evaluator_Draft4_Long_Enum_Short_Strings	9.817049245708724 ns/iter	12.740211506129116 ns/iter	0.77
Evaluator_Draft4_Type_Object	5.908043370165953 ns/iter	8.04992899014708 ns/iter	0.73
Evaluator_Draft6_Property_Names	95.60106404932046 ns/iter	125.5883808885968 ns/iter	0.76
Evaluator_Draft7_If_Then_Else	26.24418393754093 ns/iter	34.56532018390313 ns/iter	0.76
Evaluator_Draft7_Vercel_1	33014.025684843495 ns/iter	44769.56257649099 ns/iter	0.74
Evaluator_Draft7_Helm_Chart_Lock_1	111.34383905636382 ns/iter	138.13669675721914 ns/iter	0.81
Evaluator_Draft7_Helm_Chart_Lock_2	104.4284549354584 ns/iter	134.58666460428105 ns/iter	0.78
Evaluator_Draft7_Jasmine_1	106.72204712288529 ns/iter	158.3842014636993 ns/iter	0.67
Evaluator_Draft7_Dependabot_1	741.1574458192081 ns/iter	962.0183519338063 ns/iter	0.77
Evaluator_2019_09_Unevaluated_Properties	36.42114338934061 ns/iter	50.75916669999856 ns/iter	0.72
Evaluator_2019_09_OMC_JSON_V2_1	965.3233957384659 ns/iter	1339.0691159210646 ns/iter	0.72
Evaluator_2020_12_Dynamic_Ref	403.5391981624088 ns/iter	490.2068786578752 ns/iter	0.82
Evaluator_2020_12_Dynamic_Ref_Single	5.645857108685083 ns/iter	6.660412293595355 ns/iter	0.85
Evaluator_2020_12_CQL_1	2459.286241943554 ns/iter	2840.1839373391526 ns/iter	0.87
Evaluator_2020_12_OpenAPI	160409.43136761815 ns/iter	201753.37065519788 ns/iter	0.80

This comment was automatically generated by workflow using github-action-benchmark.

[github-actions]

Benchmark (macos/gcc)

Benchmark suite	Current: eedfb8e	Previous: 948ff42	Ratio
Compiler_Draft6_AdaptiveCard	60189805030.822754 ns/iter	62149394989.01367 ns/iter	0.97
Compiler_2019_09_OMC_JSON_V2	8073983192.443848 ns/iter	8439273118.972778 ns/iter	0.96
Evaluator_Draft4_Meta_1_No_Callback	159.52314298881572 ns/iter	191.23994242141546 ns/iter	0.83
Evaluator_Draft4_Required_Properties	317.47745429788074 ns/iter	373.4863770429236 ns/iter	0.85
Evaluator_Draft4_Many_Optional_Properties_Minimal_Match	18.59646750735067 ns/iter	22.393473022631248 ns/iter	0.83
Evaluator_Draft4_Few_Optional_Properties_Minimal_Match	7.204556102387926 ns/iter	9.597184809283359 ns/iter	0.75
Evaluator_Draft4_Items_Schema	439.24847774160884 ns/iter	596.4238053902826 ns/iter	0.74
Evaluator_Draft4_Nested_Object	2.049115583793565 ns/iter	2.1816040713940987 ns/iter	0.94
Evaluator_Draft4_Properties_Triad_Optional	804.2525366220336 ns/iter	889.0806348010688 ns/iter	0.90
Evaluator_Draft4_Properties_Triad_Closed	651.2549132604439 ns/iter	757.5436404547306 ns/iter	0.86
Evaluator_Draft4_Properties_Triad_Required	816.4232885838622 ns/iter	920.1548159841592 ns/iter	0.89
Evaluator_Draft4_Properties_Closed	65.67706104829438 ns/iter	87.31750577444393 ns/iter	0.75
Evaluator_Draft4_Non_Recursive_Ref	12.703967032269222 ns/iter	14.613592810920931 ns/iter	0.87
Evaluator_Draft4_Pattern_Properties_True	182.48455943293268 ns/iter	206.62383081215097 ns/iter	0.88
Evaluator_Draft4_Ref_To_Single_Property	7.0039962721272495 ns/iter	7.472020779468345 ns/iter	0.94
Evaluator_Draft4_Additional_Properties_Type	24.211500429804666 ns/iter	31.460650055828136 ns/iter	0.77
Evaluator_Draft4_Nested_Oneof	55.81623414662195 ns/iter	59.75039324431016 ns/iter	0.93
Evaluator_Draft4_Short_Enum	8.042344326322189 ns/iter	8.822358417753616 ns/iter	0.91
Evaluator_Draft4_Long_Enum	9.640942615858386 ns/iter	9.710155104240057 ns/iter	0.99
Evaluator_Draft4_Long_Enum_Short_Strings	9.07926488273047 ns/iter	8.221891930414822 ns/iter	1.10
Evaluator_Draft4_Type_Object	4.634940320346698 ns/iter	4.673648286645476 ns/iter	0.99
Evaluator_Draft6_Property_Names	106.68568439016516 ns/iter	113.89489339934102 ns/iter	0.94
Evaluator_Draft7_If_Then_Else	26.63087298551005 ns/iter	30.280366430066486 ns/iter	0.88
Evaluator_Draft7_Vercel_1	43449.95409449384 ns/iter	43821.91533090722 ns/iter	0.99
Evaluator_Draft7_Helm_Chart_Lock_1	177.82411471171258 ns/iter	178.9649020917839 ns/iter	0.99
Evaluator_Draft7_Helm_Chart_Lock_2	166.33107271825637 ns/iter	175.85674884224298 ns/iter	0.95
Evaluator_Draft7_Jasmine_1	130.95458978656782 ns/iter	145.0467125479412 ns/iter	0.90
Evaluator_Draft7_Dependabot_1	940.4802827649074 ns/iter	969.3054284113986 ns/iter	0.97
Evaluator_2019_09_Unevaluated_Properties	34.97079384929331 ns/iter	35.872648102444074 ns/iter	0.97
Evaluator_2019_09_OMC_JSON_V2_1	1044.8438865186313 ns/iter	1057.1723022736335 ns/iter	0.99
Evaluator_2020_12_Dynamic_Ref	461.9503836599327 ns/iter	448.6718664368614 ns/iter	1.03
Evaluator_2020_12_Dynamic_Ref_Single	4.810244811136937 ns/iter	4.560672313260583 ns/iter	1.05
Evaluator_2020_12_CQL_1	2526.5369819961475 ns/iter	2625.478350597879 ns/iter	0.96
Evaluator_2020_12_OpenAPI	170264.4661361096 ns/iter	276689.6236507627 ns/iter	0.62

This comment was automatically generated by workflow using github-action-benchmark.

[github-actions]

Benchmark (linux/gcc)

Benchmark suite	Current: eedfb8e	Previous: 948ff42	Ratio
Evaluator_2020_12_Dynamic_Ref	631.0818547075328 ns/iter	634.9157483528479 ns/iter	0.99
Evaluator_2020_12_Dynamic_Ref_Single	9.73029761756413 ns/iter	9.343948255419704 ns/iter	1.04
Evaluator_2020_12_CQL_1	3802.3879901334217 ns/iter	3784.995509095865 ns/iter	1.00
Evaluator_2020_12_OpenAPI	229935.04229507232 ns/iter	226326.13008130927 ns/iter	1.02
Evaluator_2019_09_Unevaluated_Properties	53.00667890430672 ns/iter	53.880314891576674 ns/iter	0.98
Evaluator_2019_09_OMC_JSON_V2_1	1362.6150602753721 ns/iter	1495.4300668669046 ns/iter	0.91
Evaluator_Draft7_If_Then_Else	45.34305472830434 ns/iter	45.1577688846125 ns/iter	1.00
Evaluator_Draft7_Vercel_1	46663.06132138789 ns/iter	47724.161990650675 ns/iter	0.98
Evaluator_Draft7_Helm_Chart_Lock_1	124.96467449970417 ns/iter	126.32508740328751 ns/iter	0.99
Evaluator_Draft7_Helm_Chart_Lock_2	123.33911417451677 ns/iter	122.72965346423626 ns/iter	1.00
Evaluator_Draft7_Jasmine_1	153.0179766353647 ns/iter	164.67718349407988 ns/iter	0.93
Evaluator_Draft7_Dependabot_1	938.8019319815683 ns/iter	935.4511028021172 ns/iter	1.00
Evaluator_Draft6_Property_Names	173.0269869881307 ns/iter	195.81746458909933 ns/iter	0.88
Evaluator_Draft4_Meta_1_No_Callback	243.80818135844282 ns/iter	243.46948881325656 ns/iter	1.00
Evaluator_Draft4_Required_Properties	460.4548663818013 ns/iter	468.8564483661102 ns/iter	0.98
Evaluator_Draft4_Many_Optional_Properties_Minimal_Match	29.676198536367934 ns/iter	29.17354595127744 ns/iter	1.02
Evaluator_Draft4_Few_Optional_Properties_Minimal_Match	10.379973321192095 ns/iter	10.283490798971632 ns/iter	1.01
Evaluator_Draft4_Items_Schema	495.63595632692244 ns/iter	507.00778581952335 ns/iter	0.98
Evaluator_Draft4_Nested_Object	4.249916951308043 ns/iter	3.8604097309468632 ns/iter	1.10
Evaluator_Draft4_Properties_Triad_Optional	1046.5859008753278 ns/iter	1048.9824534438778 ns/iter	1.00
Evaluator_Draft4_Properties_Triad_Closed	833.3751650090759 ns/iter	830.5912291246178 ns/iter	1.00
Evaluator_Draft4_Properties_Triad_Required	1032.9470871760077 ns/iter	1041.001573948017 ns/iter	0.99
Evaluator_Draft4_Properties_Closed	74.70184768516063 ns/iter	72.96655774241094 ns/iter	1.02
Evaluator_Draft4_Non_Recursive_Ref	12.536764766115382 ns/iter	12.780182721597882 ns/iter	0.98
Evaluator_Draft4_Pattern_Properties_True	289.8557108195994 ns/iter	275.1816965255089 ns/iter	1.05
Evaluator_Draft4_Ref_To_Single_Property	10.356341442038067 ns/iter	10.314375569149613 ns/iter	1.00
Evaluator_Draft4_Additional_Properties_Type	15.357552193788242 ns/iter	15.573346114262723 ns/iter	0.99
Evaluator_Draft4_Nested_Oneof	108.72135111148832 ns/iter	112.82183884708809 ns/iter	0.96
Evaluator_Draft4_Short_Enum	18.06507201825563 ns/iter	17.727768609640655 ns/iter	1.02
Evaluator_Draft4_Long_Enum	26.062955217527954 ns/iter	23.303412677018194 ns/iter	1.12
Evaluator_Draft4_Long_Enum_Short_Strings	18.060033131812276 ns/iter	17.79957501691184 ns/iter	1.01
Evaluator_Draft4_Type_Object	9.797904681908534 ns/iter	9.337990985024707 ns/iter	1.05
Compiler_2019_09_OMC_JSON_V2	10686997476.999977 ns/iter	10747970438.000038 ns/iter	0.99
Compiler_Draft6_AdaptiveCard	69797893883.99991 ns/iter	68009856280.00001 ns/iter	1.03

This comment was automatically generated by workflow using github-action-benchmark.

[github-actions]

Benchmark (windows/msvc)

Benchmark suite	Current: eedfb8e	Previous: 948ff42	Ratio
Compiler_Draft6_AdaptiveCard	168833258300 ns/iter	163445904700.00003 ns/iter	1.03
Compiler_2019_09_OMC_JSON_V2	24341284300.0001 ns/iter	23953018500.0001 ns/iter	1.02
Evaluator_Draft4_Meta_1_No_Callback	351.19128996583976 ns/iter	371.44627295597496 ns/iter	0.95
Evaluator_Draft4_Required_Properties	749.3191964286692 ns/iter	770.2389508928828 ns/iter	0.97
Evaluator_Draft4_Many_Optional_Properties_Minimal_Match	39.94605468750555 ns/iter	37.736573660713496 ns/iter	1.06
Evaluator_Draft4_Few_Optional_Properties_Minimal_Match	16.836752119395747 ns/iter	15.583415178568316 ns/iter	1.08
Evaluator_Draft4_Items_Schema	686.8388392857431 ns/iter	700.7735491070731 ns/iter	0.98
Evaluator_Draft4_Nested_Object	6.915763392857853 ns/iter	6.032035999999152 ns/iter	1.15
Evaluator_Draft4_Properties_Triad_Optional	1565.314508928495 ns/iter	1603.23482142845 ns/iter	0.98
Evaluator_Draft4_Properties_Triad_Closed	1273.360000000017 ns/iter	1309.3073214284752 ns/iter	0.97
Evaluator_Draft4_Properties_Triad_Required	1552.8819196428044 ns/iter	1597.3600446430275 ns/iter	0.97
Evaluator_Draft4_Properties_Closed	91.37478794643629 ns/iter	89.54558975241797 ns/iter	1.02
Evaluator_Draft4_Non_Recursive_Ref	35.36272206004814 ns/iter	31.9732504995821 ns/iter	1.11
Evaluator_Draft4_Pattern_Properties_True	222.02602432745567 ns/iter	229.09621327048256 ns/iter	0.97
Evaluator_Draft4_Ref_To_Single_Property	16.899213949334076 ns/iter	15.966292410714646 ns/iter	1.06
Evaluator_Draft4_Additional_Properties_Type	65.43951785713732 ns/iter	58.102008928569504 ns/iter	1.13
Evaluator_Draft4_Nested_Oneof	103.99364267885677 ns/iter	109.6426424266797 ns/iter	0.95
Evaluator_Draft4_Short_Enum	21.98785937499892 ns/iter	24.166450000003742 ns/iter	0.91
Evaluator_Draft4_Long_Enum	24.017909330156222 ns/iter	22.749354656810034 ns/iter	1.06
Evaluator_Draft4_Long_Enum_Short_Strings	24.441367857142332 ns/iter	28.53157098152246 ns/iter	0.86
Evaluator_Draft4_Type_Object	13.743899938643764 ns/iter	13.499971814728742 ns/iter	1.02
Evaluator_Draft6_Property_Names	157.81620535713716 ns/iter	166.16935267858173 ns/iter	0.95
Evaluator_Draft7_If_Then_Else	51.11156000000392 ns/iter	53.68955357141659 ns/iter	0.95
Evaluator_Draft7_Vercel_1	69605.84821429044 ns/iter	70558.23214285185 ns/iter	0.99
Evaluator_Draft7_Helm_Chart_Lock_1	296.9475740013867 ns/iter	287.7396152076016 ns/iter	1.03
Evaluator_Draft7_Helm_Chart_Lock_2	271.59743846419576 ns/iter	263.15238451574913 ns/iter	1.03
Evaluator_Draft7_Jasmine_1	217.99737500000305 ns/iter	221.1005000000199 ns/iter	0.99
Evaluator_Draft7_Dependabot_1	1513.3643913551323 ns/iter	1557.8464285712 ns/iter	0.97
Evaluator_2019_09_Unevaluated_Properties	64.29816071428515 ns/iter	75.27516741071392 ns/iter	0.85
Evaluator_2019_09_OMC_JSON_V2_1	1840.68860775753 ns/iter	1877.4745334594957 ns/iter	0.98
Evaluator_2020_12_Dynamic_Ref	861.0914905840947 ns/iter	852.3513392856127 ns/iter	1.01
Evaluator_2020_12_Dynamic_Ref_Single	13.371856413516173 ns/iter	13.35072047611205 ns/iter	1.00
Evaluator_2020_12_CQL_1	4870.283567217909 ns/iter	5070.947000001524 ns/iter	0.96
Evaluator_2020_12_OpenAPI	277118.60184819903 ns/iter	290403.73644033866 ns/iter	0.95

This comment was automatically generated by workflow using github-action-benchmark.

[michaelmior]

@jviotti Done! https://github.com/sourcemeta-research/jsonschema-benchmark/actions/runs/13546684327

[michaelmior]

Just glancing through the results, it looks like cold runs are slightly slower on average and warm runs faster. I think we'll have to do a more thorough analysis to see if this is this worthwhile.