FYI performance of hash algorithms

Path	csiph.com!usenet.pasdenom.info!weretis.net!feeder1.news.weretis.net!feeder.erje.net!news-1.dfn.de!news.dfn.de!news.informatik.hu-berlin.de!fu-berlin.de!uni-berlin.de!not-for-mail
From	"M. Strobel" <sorry_no_mail_here@nowhere.dee>
Newsgroups	comp.lang.php
Subject	FYI performance of hash algorithms
Date	Mon, 23 Apr 2012 21:42:01 +0200
Lines	57
Message-ID	<9vlpk9Fd7cU1@mid.uni-berlin.de> (permalink)
Mime-Version	1.0
Content-Type	text/plain; charset=UTF-8
Content-Transfer-Encoding	7bit
X-Trace	news.uni-berlin.de 1CeS4tleK7ghXI9aJ8jlJQOYJrTm12/vnxaIBAPVwWhGUxlhs=
User-Agent	Mozilla/5.0 (X11; U; Linux x86_64; de; rv:1.9.2.28) Gecko/20120306 Thunderbird/3.1.20
Xref	csiph.com comp.lang.php:5688

Show key headers only | View raw

I am just asking myself which of the many hash algorithms performs best, so I can use
it to index my translation strings.

So I wrote a test script. It iterates over an array of 3.917.116 strings calculating
the hash of each.

Here you have the results, with run time, and run time relative to hash length
(smaller is better).
Interesting that the tiger hashes seem to perform well.

There were only collisions with adler32 (many), crc32 and crc32b.

/Str.

Name	Length	Timing	   t/Byte (skal.)
crc32	8	2,2150330544	276,9
crc32b	8	2,2309110165	278,9
adler32	8	2,4038968086	300,5
md4	32	2,9871790409	93,3
md5	32	3,0691349506	95,9
tiger128,3 32	3,4266569614	107,1
tiger160,3 40	3,5735139847	89,3
tiger128,4 32	3,6424109936	113,8
tiger192,3 48	3,6483809948	76,0
sha1	   40	3,6727621555	91,8
tiger160,4 40	3,7693319321	94,2
tiger192,4 48	3,8536069393	80,3
ripemd128  32	3,9478549957	123,4
ripemd256  64	4,2519729137	66,4
salsa20	128	4,4366490841	34,7
salsa10	128	4,4819700718	35,0
ripemd160 40	4,8140711784	120,4
sha224	56	5,2336411476	93,5
ripemd320 80	5,2477719784	65,6
sha256	  64 	5,3834650517	84,1
haval128,3 32	5,8804600239	183,8
haval160,3 40	5,9342179298	148,4
haval192,3 48	6,0202069283	125,4
haval224,3 56	6,0966057777	108,9
haval256,3 64	6,2231030464	97,2
sha384	   96	6,4372200966	67,1
sha512	128 	6,7605450153	52,8
haval128,4 32	7,0831079483	221,3
haval160,4 40	7,1777567863	179,4
haval192,4 48	7,2292518616	150,6
haval224,4 56	7,3201220036	130,7
haval256,4 64	7,4621391296	116,6
whirlpool 128	7,4762940407	58,4
haval128,5 32	7,9410018921	248,2
haval160,5 40	8,028870821	200,7
haval192,5 48	8,102463007	168,8
haval224,5 56	8,1795060635	146,1
haval256,5 64	8,3195581436	130,0
gost	   64 	12,0381391048	188,1
snefru	   64	14,0808851719	220,0
snefru256  64	14,1807420254	221,6
md2	   32	17,4794299603	546,2

Back to comp.lang.php | Previous | Next — Next in thread | Find similar | Unroll thread

Thread

FYI performance of hash algorithms "M. Strobel" <sorry_no_mail_here@nowhere.dee> - 2012-04-23 21:42 +0200
  Re: FYI performance of hash algorithms Jerry Stuckle <jstucklex@attglobal.net> - 2012-04-23 16:16 -0400
    Re: FYI performance of hash algorithms "M. Strobel" <sorry_no_mail_here@nowhere.dee> - 2012-04-23 22:44 +0200

csiph-web