[RC5] Do we check for the Pentium bug?

Greg Wooledge wooledge at kellnet.com
Mon Jun 7 19:51:56 EDT 1999

Aaron Kramer (akramer at voicenet.com) wrote:

> With all of the turnover in the used computer market, do any of our
> clients check for the 'buggy' pentium cpu? Just a thought...

Which bug do you mean?  The FDIV bug, the f00f bug, ...?

The FDIV bug isn't relevant to the RC5/DES client because the client
doesn't do any floating point arithmetic in the core.  (It may do a
little bit when computing keys/sec, but that doesn't affect the validity
of the cryptography, and in any case the FDIV bug yields only a *small*
loss of accuracy.)

The f00f bug isn't relevant, either.  It causes the CPU to hang when a
certain invalid instruction is issued -- but the client doesn't try to
execute this instruction, so it won't trigger the bug.  (I'm running
the Linux client on such a system at work, and it hasn't crashed yet.
But then, the Linux kernel has a f00f bug workaround anyway.)

Greg Wooledge                    | Distributed.NET http://www.distributed.net/
wooledge at kellnet.com             | because a CPU is a terrible thing to waste.
http://www.kellnet.com/wooledge/ |

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