[RC5] Estimated time of completion of DES-II-2, please.

Peter A. DeNitto denitto at llamas.net
Tue Jun 23 11:01:01 EDT 1998


On Mon, 22 Jun 1998, Jason Untulis wrote:

> "Hugo Villeneuve" <Hugo_Villeneuve at videotron.ca> said:
> 
> o A more accurate estimate comes from the comparative stats page at
>   <http://www.distributed.net/statistics/stats.html> (note this hasn't
>   been updated for a week...) This gives us about 3.4x DES-to-RC5 ratio
>   (although this doesn't scale since it's based upon completed block
>   distribution by machine and reference CPUs representing an entire

I've got a preliminary stats2.html there using a PII/266 (non celeron) as
the x86 bench chip.  Some people feel this is a more representitive chip
for the x86 line, as the BrydDes library for x86 is heavily P5 optimized,
and the P100 rc5/des ratio is a bit large.

This also contains the DES speed of the new Ultra Sparc client, which
helps bridge the gap by the change in x86 bench cpu.

> o The switchover (alluded to in an earlier post). For example, all the Mac
>   people (including myself) who run the Meggs' graphical client rather than
>   the DeNitto dual core client. My guess is that we'll effectively lose at
>   least a day's worth of distributed.net's power because of this, although
>   some of it is inevitable.

Precicely.  It would be nice if we could get everyone to upgrade to a
v2.7016 or greater client, especially to the 2.71's that are planned to be
released this week.

> o Those people who left distributed.net for other projects (Golomb rulers,
>   Mersenne primes) because they think that RC5-64 is a waste of time, but
>   that the DES contests aren't and come back to add to the DES contest.
>   (This wouldn't be a problem if v3 actually made some progress in
>   increments smaller than 3 months (and growing), hint, hint.)

Yes, we can expect all those RISC machines to come back for DES II-2 like
they did for DES II-1.

> o Server I/O apparently may be a factor if a couple UW SCSI disks aren't
>   found/bought/donated. (see the latest State of the D.Net Address)

This is only a factor with regards to statistics.  Because of the proxy
network setup, clients will be able to get blocks assigned, so no work
will be wasted in that respect.  There may be some random blocks
introduced if the master server is not able to spend time issuing blocks
because it's busy writing the logs to disk. :/

--Pete

--
Help Crack Government Encryption - Join the Bovine Project: www.distributed.net
The cows are tromping all over America. Cows can replicate. They're
being born all over the world. There's plenty of beef available!
		-- US Representative Zoe Lofgren (D-Ca)

--
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