David Schwartz wrote: > On Sun, 11 Aug 2002 17:54:49 -0700 (PDT), James Shelby wrote: > > >>My first thought was the same. Which brought up >>another interesting question....the 32bit Pentium II >>333 is still faster than the UltraSparc 400. > > > Frankly, I don't find this surprising. A RISC CPU would likely be slower > than a CISC CPU at the same clock rate. As for Pentium's being a 32bit CPU, > that's just the width of their general-purpose registers. Many internal > pathways are wider than that. Bit width is not a good indicator of > performance, especially on algorithms that don't benefit much from wider > numeric types. > > >>I am running Solaris 9 in 64bit mode and used the >>default for the latest Openssl96g. I get 2,700,000 >>Des calculations per minute on the Pentium and >>2,400,00 on the UltraSparc with 32bit mode libraries. >>I have just compiled the libraries using Solaris64-gcc >>and compiled my program using -m64 and receive >>1,700,000 calculations per second. Hmm, is it just me >>or is something wrong with that math..heh... > > > That's kind of surprising. DES uses a lot of operations on numbers with bit > widths greater than 32 but less than or equal to 64. One would think it would > get a boost.
Have you checked whether there's hand-optimised assembler for the various options? Cheers, Ben. -- http://www.apache-ssl.org/ben.html http://www.thebunker.net/ Available for contract work. "There is no limit to what a man can do or how far he can go if he doesn't mind who gets the credit." - Robert Woodruff ______________________________________________________________________ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
