[RC5] Strange Keyrates

gindrup at okway.okstate.edu gindrup at okway.okstate.edu
Tue Jun 16 10:00:37 EDT 1998


     PPros on d.net cracking code probably gain the greatest 
     architectural benefit over P5s due to speculative execution and 
     branch prediction.  Since the core is a tight little loop that's 
     executed *many* times, the branch prediction is almost always 
     correct and so the predictive executions are almost always kept.  
     The biggest improvement for the PPro running the d.net client would 
     be to reduce the rate of (prediction invalidating) interrupts and 
     mode switches.  (Nuts&Bolts's WinGauge tells me I'm getting a few 
     thousand of these each second -- most of them internally produced 
     (intentional) GPFs.)
     
     PPro L2 cache may be a factor, but not much of one (the crack code 
     and data are very small for the d.net clients).  There may be 
     microscopic differences caused by the reoptimization of the 32-bit 
     integer instructions on the PPro, but later P5s (esp. MMXs) 
     incorporated some of these improvements.
     
     My PPro/200 running Win32 GUI v2.7016.400 gets around 560kk/s on an 
     idle machine.  (YMMV)  Scaling up to 266 gives ~745kk/s which is in 
     line with the observation that PIIs don't scale any better than P5s 
     and PPros (i.e, an overclocked PPro or P5 performs about as well as 
     the same speed PII.  I expect this prediction to break on the newer 
     100MHz FSB architectures, but not so much for the d.net client code 
     which performance has very little to do with bus speeds under the 
     processors mentioned).
            -- Eric Gindrup ! gindrup at Okway.okstate.edu


______________________________ Reply Separator _________________________________
Subject: Re: [RC5] Strange Keyrates 
Author:  <rc5 at llamas.net> at SMTP
Date:    6/15/98 1:01 PM


These are the rates I have gathered. My trusty Cyrix 486 100 can do about 
75kk/s if it is the only thing running. A P120/133 can handle about 235kk/s 
and a P2 266 can do 750kk/s. Are these freak figures or am I lucky?
     
It seems to me that PPro machines are better than MMX ones at doing keys, 
and P2's better than PPros. I figured that PPro machines were quicker than 
MMX because they had more math code on chip. Am I right?
     
Mike Reed, mailto:%20Michael at Reedsweb.net 
http://www.reedsweb.net -- Affordable web hosting for the UK
     
--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net 
rc5-digest subscribers replace rc5 with rc5-digest
     
     


--
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