Re: Added hashes.
Available news archives: comp.lang.tcl - comp.lang.python - comp.security.firewalls - sci.crypt - comp.lang.php - comp.lang.javascript
Google
 
Web news.hping.org


sci.crypt archive

Re: Added hashes.

From: Joseph Ashwood <ashwood@msn.com>
Date: Mon Dec 12 2005 - 09:37:49 CET

"Colin Percival" <cperciva@sfu.ca> wrote in message
news:dnj8d5$np5$1@morgoth.sfu.ca...
> Paul Rubin <http://phr.cx@nospam.invalid> wrote:
>> Ian Stirling <root@mauve.demon.co.uk> writes:
>>> I'll disagree slightly with the 'half speed' comment - as in many cases
>>> the limiting factor is disk speed, but of course it may be an issue.
>>
>> If you're talking about modern pc's and hashes like sha256, the cpu
>> won't be able to keep up with the disk even for just one hash.
>
> On an AMD Athlon XP 2400+, I can sha256 a 1.1GB file in 16.3 seconds of
> CPU time. It takes over twice that much disk time.

That's only 69MB/sec (admirable though it is), a single Seagate Cheetah can
handle up to 142MB/sec according to the spec sheet
(http://www.seagate.com/docs/pdf/marketing/ds_1566_004_cheetah_15k_4.pdf).
In high performance situations where drives will be RAIDed the throughput
will be even higher, and I've personally seen environments where 2Gbps
fibrechannel was completely filled. For a "normal" desktop 69MB/sec is
probably above the normal from the drive, but even there the difference is
shrinking.
                    Joe
Received on Fri Dec 23 20:09:34 2005