Re: Twofish perfomance?

Bruce Schneier (schneier@counterpane.com)
Thu, 18 Jun 1998 07:10:21 -0500

• Next message: Simon R Knight: "Re: Locking physical memory (RAM) under Windows"
• Previous message: William H. Geiger III: "Re: Locking physical memory (RAM) under Windows"
• In reply to: Kriston J. Rehberg: "Re: Locking physical memory (RAM) under Windows"

The paper gives performance as a relation of key size. With the
particular keying option used in the 17.8 figure, it is not affected
by key size.

RC5 is will be faster than Twofish on the Pentium Pro/II, but much
slower on the Pentium.

At 08:32 AM 6/17/98 -0700, Alex Alten wrote:
>Bruce,
>
>On your web site you claim 17.8 cycles/byte on a pentium pro.
>I'm not sure what this means since these processors do
>instruction caching/pipelining and data caching. If

We take all that into account.

>I were to use a 200 MHz PPro then this apparently translates
>to about 11 MBytes/sec. This seems slower than RC5 which
>does about 15 MBytes/sec according to a prior email by
>Eric Young. Is this accurate? BTW, does the performance
>vary by key size?
>
>- Alex
>
>--
>Alex Alten
>Andrade@Netcom.Com
>P.O. Box 11406
>Pleasanton, CA 94588 USA
>(510) 417-0159
>

**********************************************************************
Bruce Schneier, President, Counterpane Systems Phone: 612-823-1098
101 E Minnehaha Parkway, Minneapolis, MN 55419 Fax: 612-823-1590
                                            http://www.counterpane.com

• Next message: Simon R Knight: "Re: Locking physical memory (RAM) under Windows"
• Previous message: William H. Geiger III: "Re: Locking physical memory (RAM) under Windows"
• In reply to: Kriston J. Rehberg: "Re: Locking physical memory (RAM) under Windows"