[RC5] Different DES code in Irix mips clients?

Joe Zbiciak j-zbiciak1 at ti.com
Wed May 27 22:06:46 EDT 1998


'Samel Vitezslav' said previously:

| These are my benchmarks (on totally idle machine):
| 
|       build                   keys/sec
|      v2.7010.393              1173835.26
|      v2.7013.396 32 bit       1206998.96
|      v2.7013.396 64 bit       2450650.49
|      v2.7020.403               603842.14
| 
| (tested on IRIX64 - MIPS R10000 / 175MHz)
| 
|   Authors of 396 code refer to DES code as "Meggs' DES C code".
| 
|   So my question is:
|   Is there an intent to build DES code based on this faster slice ??

Are you comparing DES against DES?  The DES-II contest that started in
January is over, and so all of the clients should be cracking RC5
blocks unless you are running offline, causing the client to generate
random blocks.  If that's the case, then your clients are not doing
useful work with respect to d.net.

Unless I'm mistaken, the MIPS processors don't have a rotate
instruction, and so are at a disadvantage for RC5 as compared to CPUs
which do have a rotate instruction (eg. x86, PPC, 68k).  The number
you've quoted appears to be in the ballpark for an RC5 keyrate on a CPU
w/out a rotate instruction, rather than a DES keyrate.

If you are comparing DES against DES, then this is a problem.  I would
imagine that we will have a client built around the fastest available
core by the next DES-II contest. (Is that July 13th?  I forget.)

Regards,

--Joe

--
  +------- Joseph Zbiciak ------+
  |- - - j-zbiciak1 at ti.com - - -|   without you, everything falls apart
  | -Texas Instruments, Dallas- |   without you, it's not as much fun
  |- - #include <disclaim.h> - -|   - NIN -      to pick up the pieces. 
  +-----------------------------+
--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list