[RC5] Experimenting

Peter Cordes peter at llama.nslug.ns.ca
Tue Feb 15 21:46:46 EST 2000


On Tue, Feb 15, 2000 at 07:43:19PM -0500, EnoJon wrote:
> Actually, it's a very handy utility.  Esp when you bump the Kernel32
> up to "realtime".  This is nothing new for WinNT users, but a lacking
> tool on Win95-98.  Also, dnetc uses all available "idle" cycles, if
> you run it at pri=9 and as a "normal" or "low" priority process, you
> do not get a boost in keys/sec, but you do get more block throughput
> (ie: more packets completed per hour of real time).
> 
> It's the old virtual cpu time vs. wallclock.  Even with a fully
> loaded system, the dnetc client reports a steady benchmark rate, but
> it can mean 32 mins vs 2 hrs to complete a 32*2^28 packet.  My average
> keyrate has gone from a paltry 432K keys/sec to 500K+ keys/sec on a
> Pent MMX 200.

 WTF!?! How did you get it to do that?  The client benchmarks at 431,317.59
kk/s for me (on my P200MMX, 64MB RAM, 430TX chipset), so how can you get
500+ kk/s?

 Since the benchmark measures keys / CPU time, it should be impossible to go
any _faster_ than the benchmark, since real time is always >= CPU time.

 Sorry I can't try this under w9x, but I don't have any Micros~1 operating
systems installed on this machine.  (and my beautiful uptime... 33 days :).

 What client version are you running?  have there been any improvements to
the x86 P5/MMX core since v2.8005-453?

-- 
#define X(x,y) x##y
DUPS Secretary ; http://is2.dal.ca/~dups/
Peter Cordes ;  e-mail: X(peter at cordes.phys. , dal.ca)

"The gods confound the man who first found out how to distinguish the hours!
 Confound him, too, who in this place set up a sundial, to cut and hack
 my day so wretchedly into small pieces!" -- Plautus, 200 BCE

--
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