[RC5] RC5 on Pentium4

Jason Hartzell jhartzell at arcataassoc.com
Tue Aug 27 10:56:51 EDT 2002


Because there is nothing in the stats database to reflect the idea that the
Cow performs at such a high rate while executing under DOS when compared to
other operating systems. If your idea were in any way accurate, it would be
reflected in every processor version under the speed database. You could
look up an AMD K6-2 500Mhz, a Pentium Pro 200Mhz, and 1.4 Ghz Celeron. They
should all show the 50% gain for running DOS, but they don't. Why the change
with the 1700Mhz Pentium 4? My guess is a typo when entering the speed, or
maybe someone is trying to overclock their processor (I'm not even sure if
you can overclock a P4) and submitted the speed under the original processor
speed instead of the overclocked speed. 

My personal experience of running client under DOS show minor performance
gains, if any. Not a 50% gain over NT/2000/XP. That's really just absurd to
think you would achieve those levels.

The only thing possibly redeaming about that stat is that it is running the
2.8018 version of the client. Since that is considered a pre-release version
of the client I personally have not touched it, but the speeds database
shows significant speed gains for 2.8018 (again, not 50% over 2.8015, but it
is faster). Actually, it's a never was released version of the client. The
Windows/DOS/Linux versions of the pre-release client moved to 2.8019, while
the full release clients are still on 2.8015. I'm ready for one of those to
make it up to be a full release since version 2.8015 can't seem to identify
any of my new processors and has to run the benchmark every time it starts.



-----Original Message-----
From: waldo kitty [mailto:wkitty42 at alltel.net]
Sent: Tuesday, August 27, 2002 5:15 AM
To: rc5 at lists.distributed.net
Subject: Re: [RC5] RC5 on Pentium4



why do you think that DOS stat should be removed? obviously that system
wasn't hobbled by running any graphical shell or extraneous drivers and
stuff... pure DOS is the only system that gives full processing power of
the CPU to the single task that is running... the CPU doesn't have to
spend time doing anything else like serviceing a GUI, network layers,
mouse drivers, software based printers or modems, etc...

ANY other system is going to have to spend time doing other things...
there is simply no other way for a multitasking/multiswitching multiuser
system to work... no switching away means full power to the program
running...

Jason Hartzell wrote:
> 
> I have no idea what Bruce Fords keyrate calculator is, but...
> 
> You can always submit your own speeds to the database. Also, when you
> compare speeds, look for the same processor, OS and client revision. Those
> numbers should be pretty close.
> 
> I don't have the cow running on any P4s, but my 1400 PIII benchmarked RC5
at
> around 3.9 Mkeys/s on version 2.8015 of the client.
> 
> Also, the benchmarks I see submitted at http://www.distributed.net/speed
> show the P4-1700 to do around 2.4 MKeys/s. One person claims their P4-1700
> got 3.4 while running DOS. Probably should have that stat removed from the
> speeds database, since it looks to be WAY off.
> 
> You're actual rate will always be lower than the benchmark because your
> computer actually does work most of the time (at least it should). That's
my
> experience, and I see it as a non-issue. Maybe the benchmark runs at a
high
> priority, if you're worried about it, set the client to run at a higher
> priority and then see what numbers you get.
> 
> -----Original Message-----
> From: Mark Renfer [mailto:mri at netsec.ch]
> Sent: Sunday, August 25, 2002 9:42 AM
> To: rc5 at lists.distributed.net
> Subject: [RC5] RC5 on Pentium4
> 
> Howdy folks,
> 
> The real cracking speed and the cracking speed reported
> by the benchmark are still very different on P4 machines with
> the latest clients.
> 
> According to Bruce Fords keyrate calculator, a P4-1700
> should have a cracking speed of 3.2 MKeys/s. The benachmark
> reports 3.28 MKeys/s. So far so good. But when the client
> actually works, it does just ~2.5 MKeys/s. This has been
> double-checked, i.e. the client itself reports that low speed
> as well as our presonal proxy analyzer program.
> 
> I have already reported this behaviour last year.
> 
> Does anyone get the high cracking rate reported by the
> benchmark when the client works on real blocks?
> (Run the client in a console window until the first packet has
> been completed then check the keyrate on the "Summary"
> line)
> 
> Kind regards,       Mark
> 
> --
> 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

-- 
       _\/
      (@@)                      Waldo Kitty, Waldo's Place USA
__ooO_( )_Ooo_____________________ telnet://bbs.wpusa.dynip.com
_|_____|_____|_____|_____|_____|_____ http://www.wpusa.dynip.com
____|_____|_____|_____|_____|_____|_____ ftp://ftp.wpusa.dynip.com
_|_Eat_SPAM_to_email_me!_YUM!__|_____|_____ wkitty42 (at) alltel.net


--
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