[RC5] blocksize suggestion

Scott Wasson swasson at arrowpoint.com
Wed Jan 13 14:34:07 EST 1999


>Steve Bird (sbird at UATC.com) wrote:
>Blocksize is irrelevant. I'm leaving mine at 2^31. I'll set my buffer to
>handle as many blocks as I can compute in 1 hour and let it roll. SO
>technically, if I do find the key, I could cause d.net to waste its time
>for an hour. But, to be perfectly honest, I don't think they'll mind.

The main concern is keeping the system busy for ~12 hours (I dial in about
twice a day), without hoarding blocks.  If the preferred blocksize is set
to 31, and I set input buffers to 20, then I *might* get 8*20=160 blocks,
right?  That would be just about right for my system without excessive
hoarding.

The problem is, though, it's only a preferred blocksize.  I might get as
many as 160 blocks.  But then again I might get all 2^28 blocks, for a
total of only 20 blocks.  :-(  You just don't know what you're going to get.

To circumvent this, I do one of two things, depending on which test is
currently running:

1. When RC5 is running, I set preferred blocksize to 31, and RC5 input
blocks to 100.  There's no way of running out of work any time soon, and
hoarding RC5 blocks is a non-issue.

2. When I hear that a DES contest is about to start, I reduce the preferred
blocksize to 28 so that I get a very deterministic number of DES blocks.  I
have DES input set to 160, which I know takes about 10 hours to process.
This way I'm not forced to hoard excessively in order to guarantee that I
stay busy.

It just would be nice if this reduction to 28 could happen automatically.

-Scott

-Scott Wasson, ArrowPoint Communications
-swasson at arrowpoint.com
-(978) 692-5875 x530

--
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