Add the first 31 bytes to the hash of long strings

[michaelmior]

This adds the first 31 bytes of the string into the hash value when we are dealing with longer strings.
This should serve to reduce hash collisions and consequently, also eliminate some string comparisons.

[github-actions]

Benchmark (linux/llvm)

Benchmark suite	Current: e9b7de0	Previous: 0b28a37	Ratio
Regex_Lower_S_Or_Upper_S_Asterisk	2.5233282419440437 ns/iter	2.4910785893845415 ns/iter	1.01
Regex_Caret_Lower_S_Or_Upper_S_Asterisk_Dollar	2.4885180968588023 ns/iter	2.48607795950712 ns/iter	1.00
Regex_Period_Asterisk	2.48705779529052 ns/iter	2.4862670526578854 ns/iter	1.00
Regex_Group_Period_Asterisk_Group	2.4981381838400027 ns/iter	2.485916213443055 ns/iter	1.00
Regex_Period_Plus	2.6444819053960837 ns/iter	2.797628036102881 ns/iter	0.95
Regex_Period	2.5540249858449817 ns/iter	2.5921452362218123 ns/iter	0.99
Regex_Caret_Period_Plus_Dollar	2.4891388176003213 ns/iter	2.4932322079235614 ns/iter	1.00
Regex_Caret_Group_Period_Plus_Group_Dollar	2.4896995561869018 ns/iter	2.50495197607764 ns/iter	0.99
Regex_Caret_Period_Asterisk_Dollar	2.804597299584947 ns/iter	2.4895826186042216 ns/iter	1.13
Regex_Caret_Group_Period_Asterisk_Group_Dollar	2.673389708441055 ns/iter	2.4876284924176426 ns/iter	1.07
Regex_Caret_X_Hyphen	13.047621188638226 ns/iter	12.430354946836882 ns/iter	1.05
Regex_Period_Md_Dollar	80.50387524190218 ns/iter	77.94855197204875 ns/iter	1.03
Regex_Caret_Slash_Period_Asterisk	6.213178274982546 ns/iter	5.907134045555883 ns/iter	1.05
Regex_Caret_Period_Range_Dollar	4.0383079611906405 ns/iter	2.7985861993410044 ns/iter	1.44
Regex_Nested_Backtrack	482.85801044076993 ns/iter	478.5051068073427 ns/iter	1.01
JSON_Array_Of_Objects_Unique	379.37522011838917 ns/iter	445.58510813576174 ns/iter	0.85
JSON_Parse_1	30155.769436189898 ns/iter	30234.362335749895 ns/iter	1.00
JSON_Fast_Hash_Helm_Chart_Lock	56.05908027250021 ns/iter	56.09318835114465 ns/iter	1.00
JSON_Equality_Helm_Chart_Lock	150.39043980409355 ns/iter	144.59651741756798 ns/iter	1.04
JSON_String_Equal/10	5.612054952589856 ns/iter	5.59878452518341 ns/iter	1.00
JSON_String_Equal/100	6.228808751046261 ns/iter	6.223085107698479 ns/iter	1.00
JSON_String_Equal_Small_By_Perfect_Hash/10	0.9338000786473448 ns/iter	0.9338403213472747 ns/iter	1.00
JSON_String_Equal_Small_By_Runtime_Perfect_Hash/10	14.600803942156496 ns/iter	10.256245878042291 ns/iter	1.42
JSON_String_Fast_Hash/10	2.175493319096277 ns/iter	2.1777712277251395 ns/iter	1.00
JSON_String_Fast_Hash/100	2.176857208848693 ns/iter	2.1760528164064423 ns/iter	1.00
JSON_String_Key_Hash/10	2.6465982056691626 ns/iter	2.181056726642779 ns/iter	1.21
JSON_String_Key_Hash/100	9.006640076805082 ns/iter	2.1777565982178424 ns/iter	4.14
JSON_Object_Defines_Miss_Same_Length	3.731417907207485 ns/iter	3.7366882240936183 ns/iter	1.00
JSON_Object_Defines_Miss_Too_Small	3.7347343333141954 ns/iter	3.860911583027837 ns/iter	0.97
JSON_Object_Defines_Miss_Too_Large	3.7437524924582766 ns/iter	3.7352029552533907 ns/iter	1.00
Pointer_Object_Traverse	42.669202781111466 ns/iter	45.07719834460821 ns/iter	0.95
Pointer_Object_Try_Traverse	53.33936692005604 ns/iter	52.29117143279369 ns/iter	1.02
Pointer_Push_Back_Pointer_To_Weak_Pointer	289.4057931211268 ns/iter	345.2985636273405 ns/iter	0.84
Schema_Frame_OMC_Instances	219736706.00001088 ns/iter	228248956.66667545 ns/iter	0.96
Schema_Frame_OMC_References	144369359.79999816 ns/iter	147591740.4000029 ns/iter	0.98
Schema_Frame_OMC_Locations	140743948.9999888 ns/iter	144102830.39999285 ns/iter	0.98
Schema_Bundle_Meta_2020_12	6808712.59222933 ns/iter	6913506.594063358 ns/iter	0.98

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

[github-actions]

Benchmark (macos/llvm)

Benchmark suite	Current: e9b7de0	Previous: 0b28a37	Ratio
Regex_Lower_S_Or_Upper_S_Asterisk	1.6802609044060528 ns/iter	1.971468814589948 ns/iter	0.85
Regex_Caret_Lower_S_Or_Upper_S_Asterisk_Dollar	1.570430473414605 ns/iter	1.9810383069129447 ns/iter	0.79
Regex_Period_Asterisk	1.5818243229670417 ns/iter	1.9030525153890745 ns/iter	0.83
Regex_Group_Period_Asterisk_Group	1.620905522925156 ns/iter	1.7754747876742962 ns/iter	0.91
Regex_Period_Plus	1.9026030921283879 ns/iter	2.1150481861423245 ns/iter	0.90
Regex_Period	1.9078573316442455 ns/iter	2.0914982812789304 ns/iter	0.91
Regex_Caret_Period_Plus_Dollar	1.9298137873464398 ns/iter	2.1996536937894144 ns/iter	0.88
Regex_Caret_Group_Period_Plus_Group_Dollar	1.9087135179576717 ns/iter	2.34280539841617 ns/iter	0.81
Regex_Caret_Period_Asterisk_Dollar	1.5901313525648024 ns/iter	2.127459002801557 ns/iter	0.75
Regex_Caret_Group_Period_Asterisk_Group_Dollar	1.6071150551336124 ns/iter	1.8517843652812784 ns/iter	0.87
Regex_Caret_X_Hyphen	6.526762006653691 ns/iter	8.138388492760289 ns/iter	0.80
Regex_Period_Md_Dollar	72.76575455581576 ns/iter	82.13083841385229 ns/iter	0.89
Regex_Caret_Slash_Period_Asterisk	7.093988494350542 ns/iter	5.703814949018895 ns/iter	1.24
Regex_Caret_Period_Range_Dollar	2.0885723688979345 ns/iter	2.468873671948882 ns/iter	0.85
Regex_Nested_Backtrack	751.8536392509843 ns/iter	826.2100105965349 ns/iter	0.91
JSON_Array_Of_Objects_Unique	352.7608588411329 ns/iter	396.6354510857502 ns/iter	0.89
JSON_Parse_1	21839.104091998972 ns/iter	25373.7739427536 ns/iter	0.86
JSON_Fast_Hash_Helm_Chart_Lock	51.08242500000415 ns/iter	54.897209123168494 ns/iter	0.93
JSON_Equality_Helm_Chart_Lock	140.91325203383576 ns/iter	137.09184845005993 ns/iter	1.03
JSON_String_Equal/10	7.674078769226882 ns/iter	9.136887985854056 ns/iter	0.84
JSON_String_Equal/100	6.510334563960384 ns/iter	8.93701435944688 ns/iter	0.73
JSON_String_Equal_Small_By_Perfect_Hash/10	0.3445794855628239 ns/iter	0.3596164879326734 ns/iter	0.96
JSON_String_Equal_Small_By_Runtime_Perfect_Hash/10	3.360179323907158 ns/iter	3.112259425550567 ns/iter	1.08
JSON_String_Fast_Hash/10	1.6402432595742584 ns/iter	1.6933534129457255 ns/iter	0.97
JSON_String_Fast_Hash/100	2.220010371228592 ns/iter	2.0199947223777177 ns/iter	1.10
JSON_String_Key_Hash/10	1.5946389702495034 ns/iter	1.3494421213156904 ns/iter	1.18
JSON_String_Key_Hash/100	2.8406760329914174 ns/iter	1.3554899892753218 ns/iter	2.10
JSON_Object_Defines_Miss_Same_Length	2.761631813924041 ns/iter	2.3554787763793184 ns/iter	1.17
JSON_Object_Defines_Miss_Too_Small	3.020139244250804 ns/iter	2.356919465694408 ns/iter	1.28
JSON_Object_Defines_Miss_Too_Large	2.8643525372923833 ns/iter	2.3505822721512897 ns/iter	1.22
Pointer_Object_Traverse	16.853094519814125 ns/iter	16.47863723282853 ns/iter	1.02
Pointer_Object_Try_Traverse	22.191236371772767 ns/iter	26.000532072665646 ns/iter	0.85
Pointer_Push_Back_Pointer_To_Weak_Pointer	203.9546652303022 ns/iter	182.21521769842738 ns/iter	1.12
Schema_Frame_OMC_Instances	181989468.75000387 ns/iter	182654937.49998996 ns/iter	1.00
Schema_Frame_OMC_References	110309166.6666766 ns/iter	111103340.33334841 ns/iter	0.99
Schema_Frame_OMC_Locations	107653833.33333981 ns/iter	109743047.71428122 ns/iter	0.98
Schema_Bundle_Meta_2020_12	5345222.546105828 ns/iter	4994945.937503479 ns/iter	1.07

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

[github-actions]

Benchmark (windows/msvc)

Benchmark suite	Current: e9b7de0	Previous: 0b28a37	Ratio
Regex_Lower_S_Or_Upper_S_Asterisk	6.629631696427449 ns/iter	6.801843749999798 ns/iter	0.97
Regex_Caret_Lower_S_Or_Upper_S_Asterisk_Dollar	6.841716517856764 ns/iter	6.551072544644296 ns/iter	1.04
Regex_Period_Asterisk	6.738819196428888 ns/iter	6.781790178571613 ns/iter	0.99
Regex_Group_Period_Asterisk_Group	6.5189776785713 ns/iter	6.630018973213327 ns/iter	0.98
Regex_Period_Plus	7.096386607142132 ns/iter	7.004220982143039 ns/iter	1.01
Regex_Period	7.131385044643375 ns/iter	7.025774553570991 ns/iter	1.02
Regex_Caret_Period_Plus_Dollar	6.958386607143398 ns/iter	7.300929878121332 ns/iter	0.95
Regex_Caret_Group_Period_Plus_Group_Dollar	6.983647321428674 ns/iter	6.985732142855536 ns/iter	1.00
Regex_Caret_Period_Asterisk_Dollar	6.5905133928583535 ns/iter	6.60431875000102 ns/iter	1.00
Regex_Caret_Group_Period_Asterisk_Group_Dollar	6.781005580356642 ns/iter	6.756453571426846 ns/iter	1.00
Regex_Caret_X_Hyphen	11.969312499999774 ns/iter	11.827921874999703 ns/iter	1.01
Regex_Period_Md_Dollar	158.64817357020775 ns/iter	149.7449665292567 ns/iter	1.06
Regex_Caret_Slash_Period_Asterisk	10.047473437499121 ns/iter	10.087596875003157 ns/iter	1.00
Regex_Caret_Period_Range_Dollar	7.2061348214284795 ns/iter	7.574820312500817 ns/iter	0.95
Regex_Nested_Backtrack	600.4646000000093 ns/iter	601.6718750000142 ns/iter	1.00
JSON_Array_Of_Objects_Unique	447.8035095168997 ns/iter	444.8828607456375 ns/iter	1.01
JSON_Parse_1	79983.13914556357 ns/iter	97265.0781250195 ns/iter	0.82
JSON_Fast_Hash_Helm_Chart_Lock	66.83446428570961 ns/iter	62.85901785712359 ns/iter	1.06
JSON_Equality_Helm_Chart_Lock	187.35454074308592 ns/iter	196.9501782990953 ns/iter	0.95
JSON_String_Equal/10	9.732101742266572 ns/iter	9.640528081960102 ns/iter	1.01
JSON_String_Equal/100	9.926594152112882 ns/iter	9.968141473356747 ns/iter	1.00
JSON_String_Equal_Small_By_Perfect_Hash/10	2.1673906250001096 ns/iter	2.1656012499995825 ns/iter	1.00
JSON_String_Equal_Small_By_Runtime_Perfect_Hash/10	13.949934374998918 ns/iter	14.765575892857287 ns/iter	0.94
JSON_String_Fast_Hash/10	3.7099869576604783 ns/iter	3.7146374933667157 ns/iter	1.00
JSON_String_Fast_Hash/100	3.7172126719330736 ns/iter	3.717764145977028 ns/iter	1.00
JSON_String_Key_Hash/10	5.924887499999849 ns/iter	7.914186125383125 ns/iter	0.75
JSON_String_Key_Hash/100	19.573467672860914 ns/iter	3.713363564797977 ns/iter	5.27
JSON_Object_Defines_Miss_Same_Length	4.0304573286255625 ns/iter	4.026564292904521 ns/iter	1.00
JSON_Object_Defines_Miss_Too_Small	3.4145006303349965 ns/iter	3.4056191457041822 ns/iter	1.00
JSON_Object_Defines_Miss_Too_Large	3.722427293291128 ns/iter	3.7235582977531845 ns/iter	1.00
Pointer_Object_Traverse	50.15169999999216 ns/iter	49.78550999999243 ns/iter	1.01
Pointer_Object_Try_Traverse	74.40334821428505 ns/iter	74.35649107143101 ns/iter	1.00
Pointer_Push_Back_Pointer_To_Weak_Pointer	185.97949874817763 ns/iter	184.0254271451237 ns/iter	1.01
Schema_Frame_OMC_Instances	492983550.0000195 ns/iter	498310449.9999626 ns/iter	0.99
Schema_Frame_OMC_References	356446900.0000372 ns/iter	356110000.0000579 ns/iter	1.00
Schema_Frame_OMC_Locations	348077149.99999464 ns/iter	349000300.000057 ns/iter	1.00
Schema_Bundle_Meta_2020_12	16519014.634149598 ns/iter	16785731.707342368 ns/iter	0.98

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

[michaelmior]

Note that hashing long strings is significantly slower. This is expected, but I don't think it's a problem. We're sacrificing a bit more time to hash the strings to avoid collisions in the future. Since collisions always mean a string comparison of longer strings, I think this will be a significant win overall. We may want another microbenchmark to capture this.

[github-actions]

Benchmark (linux/gcc)

Benchmark suite	Current: e9b7de0	Previous: 0b28a37	Ratio
Schema_Frame_OMC_Instances	218018857.6666827 ns/iter	217270288.33333862 ns/iter	1.00
Schema_Frame_OMC_References	145322375.00000066 ns/iter	142297907.19996345 ns/iter	1.02
Schema_Frame_OMC_Locations	140721325.9999935 ns/iter	138847643.400004 ns/iter	1.01
Schema_Bundle_Meta_2020_12	6959065.68999817 ns/iter	6950113.000002602 ns/iter	1.00
Pointer_Object_Traverse	51.05869189999339 ns/iter	47.033963627059634 ns/iter	1.09
Pointer_Object_Try_Traverse	23.00231598285041 ns/iter	23.665808391184765 ns/iter	0.97
Pointer_Push_Back_Pointer_To_Weak_Pointer	162.38567316560997 ns/iter	160.40902295579681 ns/iter	1.01
JSON_Array_Of_Objects_Unique	379.83700041329496 ns/iter	418.21238949120243 ns/iter	0.91
JSON_Parse_1	33241.52224343277 ns/iter	33333.65759399773 ns/iter	1.00
JSON_Fast_Hash_Helm_Chart_Lock	63.176550307213574 ns/iter	68.99515777824124 ns/iter	0.92
JSON_Equality_Helm_Chart_Lock	154.79754387086635 ns/iter	146.03394645427124 ns/iter	1.06
JSON_String_Equal/10	6.343034774026637 ns/iter	6.370570242993082 ns/iter	1.00
JSON_String_Equal/100	6.96671862722581 ns/iter	6.888116965505517 ns/iter	1.01
JSON_String_Equal_Small_By_Perfect_Hash/10	0.6253404471616595 ns/iter	0.6260260465208398 ns/iter	1.00
JSON_String_Equal_Small_By_Runtime_Perfect_Hash/10	25.798365432430707 ns/iter	11.70870907184515 ns/iter	2.20
JSON_String_Fast_Hash/10	0.9342739047880316 ns/iter	0.9325681612871874 ns/iter	1.00
JSON_String_Fast_Hash/100	0.9349506775042222 ns/iter	0.9343773437464205 ns/iter	1.00
JSON_String_Key_Hash/10	1.2477170017138588 ns/iter	1.2444289295954816 ns/iter	1.00
JSON_String_Key_Hash/100	12.428112422532049 ns/iter	0.9349503801644976 ns/iter	13.29
JSON_Object_Defines_Miss_Same_Length	2.798595110741571 ns/iter	2.4907903126502777 ns/iter	1.12
JSON_Object_Defines_Miss_Too_Small	2.7964974952163897 ns/iter	2.5036917727204835 ns/iter	1.12
JSON_Object_Defines_Miss_Too_Large	2.7972156813425904 ns/iter	2.8018532421117706 ns/iter	1.00
Regex_Lower_S_Or_Upper_S_Asterisk	3.1139098965282335 ns/iter	3.114292645735623 ns/iter	1.00
Regex_Caret_Lower_S_Or_Upper_S_Asterisk_Dollar	3.108506569661275 ns/iter	2.8010424635481854 ns/iter	1.11
Regex_Period_Asterisk	2.8025741934992205 ns/iter	2.8017468996157557 ns/iter	1.00
Regex_Group_Period_Asterisk_Group	2.8006208330536895 ns/iter	2.7978297395751763 ns/iter	1.00
Regex_Period_Plus	3.114798067699511 ns/iter	3.1166811815614546 ns/iter	1.00
Regex_Period	3.1130706055271897 ns/iter	3.1090840621181366 ns/iter	1.00
Regex_Caret_Period_Plus_Dollar	3.116355738196859 ns/iter	3.1163956582886456 ns/iter	1.00
Regex_Caret_Group_Period_Plus_Group_Dollar	3.421436684120717 ns/iter	3.4183233616475555 ns/iter	1.00
Regex_Caret_Period_Asterisk_Dollar	4.042354681472904 ns/iter	3.106534319392403 ns/iter	1.30
Regex_Caret_Group_Period_Asterisk_Group_Dollar	4.5374145046367715 ns/iter	3.107237938848794 ns/iter	1.46
Regex_Caret_X_Hyphen	13.680902454681064 ns/iter	13.691540037160562 ns/iter	1.00
Regex_Period_Md_Dollar	94.3532918242415 ns/iter	89.04632832078678 ns/iter	1.06
Regex_Caret_Slash_Period_Asterisk	9.668832899762965 ns/iter	8.702872309623526 ns/iter	1.11
Regex_Caret_Period_Range_Dollar	4.381202431253075 ns/iter	3.5284094938454373 ns/iter	1.24
Regex_Nested_Backtrack	847.4524100096965 ns/iter	821.3496328165464 ns/iter	1.03

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

[github-actions]

Benchmark (macos/gcc)

Benchmark suite	Current: e9b7de0	Previous: 0b28a37	Ratio
Regex_Lower_S_Or_Upper_S_Asterisk	2.0258614407378492 ns/iter	2.0174201755549186 ns/iter	1.00
Regex_Caret_Lower_S_Or_Upper_S_Asterisk_Dollar	2.1504282606507856 ns/iter	2.039836000301083 ns/iter	1.05
Regex_Period_Asterisk	2.332995432295689 ns/iter	2.3725217817191564 ns/iter	0.98
Regex_Group_Period_Asterisk_Group	2.0444805300728444 ns/iter	2.012895502797073 ns/iter	1.02
Regex_Period_Plus	1.706666424598411 ns/iter	1.6929272298751545 ns/iter	1.01
Regex_Period	1.6925614648334848 ns/iter	1.6988639358331692 ns/iter	1.00
Regex_Caret_Period_Plus_Dollar	1.697369275779185 ns/iter	1.8308002716998528 ns/iter	0.93
Regex_Caret_Group_Period_Plus_Group_Dollar	1.7337273937204367 ns/iter	1.7141456203241394 ns/iter	1.01
Regex_Caret_Period_Asterisk_Dollar	2.025433251726799 ns/iter	2.064772111697076 ns/iter	0.98
Regex_Caret_Group_Period_Asterisk_Group_Dollar	2.0224826441718675 ns/iter	2.014473499313502 ns/iter	1.00
Regex_Caret_X_Hyphen	6.388649486379174 ns/iter	6.378212473081122 ns/iter	1.00
Regex_Period_Md_Dollar	74.41796272171298 ns/iter	76.5031822045315 ns/iter	0.97
Regex_Caret_Slash_Period_Asterisk	4.727659937424794 ns/iter	4.506638906967723 ns/iter	1.05
Regex_Caret_Period_Range_Dollar	2.056094748116322 ns/iter	2.095368903844674 ns/iter	0.98
Regex_Nested_Backtrack	850.6357505666643 ns/iter	910.6886823568311 ns/iter	0.93
JSON_Array_Of_Objects_Unique	205.05484926816038 ns/iter	239.6572223237621 ns/iter	0.86
JSON_Parse_1	23372.507722116672 ns/iter	24755.050659718545 ns/iter	0.94
JSON_Fast_Hash_Helm_Chart_Lock	24.89950726510181 ns/iter	25.145402685685355 ns/iter	0.99
JSON_Equality_Helm_Chart_Lock	124.38195161110708 ns/iter	137.41790311229246 ns/iter	0.91
JSON_String_Equal/10	5.841370681715222 ns/iter	5.9626029548846535 ns/iter	0.98
JSON_String_Equal/100	5.281034937364712 ns/iter	5.526829970671494 ns/iter	0.96
JSON_String_Equal_Small_By_Perfect_Hash/10	0.7826440687245055 ns/iter	0.8154362658001352 ns/iter	0.96
JSON_String_Equal_Small_By_Runtime_Perfect_Hash/10	3.6943804841199177 ns/iter	3.7529985880388947 ns/iter	0.98
JSON_String_Fast_Hash/10	1.9631479373444969 ns/iter	2.017165036434586 ns/iter	0.97
JSON_String_Fast_Hash/100	1.9934926678186193 ns/iter	2.0288087672718533 ns/iter	0.98
JSON_String_Key_Hash/10	1.4467364766134099 ns/iter	1.4973768876085525 ns/iter	0.97
JSON_String_Key_Hash/100	2.354123216347798 ns/iter	2.0566548370704245 ns/iter	1.14
JSON_Object_Defines_Miss_Same_Length	1.9017333871554978 ns/iter	1.9654819239370527 ns/iter	0.97
JSON_Object_Defines_Miss_Too_Small	1.8624833145553212 ns/iter	2.1618056149784173 ns/iter	0.86
JSON_Object_Defines_Miss_Too_Large	1.8837944705038694 ns/iter	1.9679741197941365 ns/iter	0.96
Pointer_Object_Traverse	55.09190808452208 ns/iter	56.54213354222707 ns/iter	0.97
Pointer_Object_Try_Traverse	39.64399314143694 ns/iter	38.987513829760296 ns/iter	1.02
Pointer_Push_Back_Pointer_To_Weak_Pointer	167.52919392418704 ns/iter	162.00948046560754 ns/iter	1.03
Schema_Frame_OMC_Instances	166051149.36828613 ns/iter	180894970.89385986 ns/iter	0.92
Schema_Frame_OMC_References	100642102.10527693 ns/iter	119090636.57124837 ns/iter	0.85
Schema_Frame_OMC_Locations	100752149.30943081 ns/iter	114684859.91160075 ns/iter	0.88
Schema_Bundle_Meta_2020_12	5040670.434633891 ns/iter	6257280.707359314 ns/iter	0.81

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

[jviotti]

Let me open a Blaze PR with this so we can see the GoogleBenchmark impact and also on the final benchmark

[jviotti]

See sourcemeta/blaze#381

[jviotti]

Making this a draft for now so we don't accidentally merge!