Re: Hash Function: MD5 or other? - Mailing list pgsql-general

From Shelby Cain
Subject Re: Hash Function: MD5 or other?
Date
Msg-id 20050614141544.22083.qmail@web50102.mail.yahoo.com
Whole thread Raw
In response to Re: Hash Function: MD5 or other?  (Greg Stark <gsstark@mit.edu>)
List pgsql-general
--- Greg Stark <gsstark@mit.edu> wrote:
>
> Note that MD5 is slow and CPU-intensive. By design.
>
> If you want a quick way to find matching records then you might find
> something
> like CRC to be more useful. With MD5 it's supposed to be hard for
> someone to
> come up with inputs that hash to a target value, but if you're not
> too worried
> about people trying to do that then MD5 is probably overkill.
>

$ ./hash -b

CRC32:      302.78 MB/sec
HAVAL 128:  165.33 MB/sec
HAVAL 160:  178.69 MB/sec
HAVAL 192:  124.74 MB/sec
HAVAL 224:  123.05 MB/sec
HAVAL 256:  98.14 MB/sec
MD2:        9.03 MB/sec
MD4:        233.36 MB/sec
MD5:        105.39 MB/sec
Panama:     311.21 MB/sec
RIPEMD-128: 129.88 MB/sec
RIPEMD-160: 76.75 MB/sec
SHA1:       135.40 MB/sec
SHA256:     49.42 MB/sec
SHA384:     32.77 MB/sec
SHA512:     31.58 MB/sec
Tiger:      54.02 MB/sec
Whirlpool:  17.51 MB/sec

Elapsed time:       3.56 seconds
Average throughput: 121.06 MB/s

Granted, MD5 isn't the quickest hashing algorithm out there but it is
certainly fast enough for general use IMO.

Regards,

Shelby Cain




__________________________________
Discover Yahoo!
Find restaurants, movies, travel and more fun for the weekend. Check it out!
http://discover.yahoo.com/weekend.html


pgsql-general by date:

Previous
From: Alex Stapleton
Date:
Subject: Re: Hash Function: MD5 or other?
Next
From: David Siebert
Date:
Subject: Re: pg_dumpall not working?