REPORT WU-SPEEDS HERE !

Thierry Van Driessche
Thierry Van Dri...
Joined: 9 Feb 05
Posts: 210
Credit: 229929
RAC: 0

> Are the semprons so good or

Message 5776 in response to message 5774

> Are the semprons so good or are the P4`s so bad ?
>
> now that I oc`d the 2200+ to 1800mhz, my time is 6:34 hrs per WU for einstein
> and 3.75 for S@H

Well, don't underestimate the P4 HT:

3.75h. for S@H means you are doing some 6.4WU's a day.

With my P4 HT , I'm doing real time some 10 to 11 WU's a day ;)

Greetings from Belgium
Thierry

ErichZann
ErichZann
Joined: 11 Feb 05
Posts: 120
Credit: 81582
RAC: 0

> > > > I was always under

Message 5777 in response to message 5775


> >
>
> I was always under the assumption that sempron's were worse than an Athlon XP.
> But my Athlon XP 2400+ takes almost 12 hours for a WU.
>

Yes, Athlon XP's are normaly faster... and as i said i need 7 hours for a e@h WU with my 2200+... something doesnt seem to be alright with yours

Jordan Wilberding
Jordan Wilberding
Joined: 19 Feb 05
Posts: 162
Credit: 715454
RAC: 0

> > > > > > > > I was

Message 5778 in response to message 5777

>
> > >
> >
> > I was always under the assumption that sempron's were worse than an
> Athlon XP.
> > But my Athlon XP 2400+ takes almost 12 hours for a WU.
> >
>
> Yes, Athlon XP's are normaly faster... and as i said i need 7 hours for a e@h
> WU with my 2200+... something doesnt seem to be alright with yours
>

Hmmm, well I have several 2400+ machines, and they are all around this speed. I wonder what it could be?

such things just should not be writ so please destroy this if you wish to live 'tis better in ignorance to dwell than to go screaming into the abyss worse than hell

Seti-Cruncher
Seti-Cruncher
Joined: 9 Feb 05
Posts: 70
Credit: 7114
RAC: 0

> I was always under the

Message 5779 in response to message 5775

> I was always under the assumption that sempron's were worse than an Athlon XP.
> But my Athlon XP 2400+ takes almost 12 hours for a WU.

That's because you are running Linux. The Einstein client performs very badly in Linux. In SETI, there is little, if any, difference between Windows and Linux.

Be lucky,

Neil

BlackAdder
BlackAdder
Joined: 9 Feb 05
Posts: 2
Credit: 561686
RAC: 0

I'm running a XP3200, Asus

I'm running a XP3200, Asus A7N8X delux MB, 1.5 gig Corsair DDR 400 ram, nothing overclocked. My Einstein times are 5:50 to 6:00 hours, seti Boinc 2:30, Mfold 1.25 55 minutes average.

KWSN_Dagger
KWSN_Dagger
Joined: 22 Jan 05
Posts: 17
Credit: 5180
RAC: 0

AMD64 3200+ with Win ME:

AMD64 3200+ with Win ME: SETI 2.5 to 3 hrs, Einstein 4 to 5 hrs, Predictor 40 to 50 mins. LHC 4 seconds to 2 hrs.

Jordan Wilberding
Jordan Wilberding
Joined: 19 Feb 05
Posts: 162
Credit: 715454
RAC: 0

> > I was always under the

Message 5782 in response to message 5779

> > I was always under the assumption that sempron's were worse than an
> Athlon XP.
> > But my Athlon XP 2400+ takes almost 12 hours for a WU.
>
> That's because you are running Linux. The Einstein client performs very badly
> in Linux. In SETI, there is little, if any, difference between Windows and
> Linux.

Why is this? You would think in Linux they could make it even better than for Windows.

such things just should not be writ so please destroy this if you wish to live 'tis better in ignorance to dwell than to go screaming into the abyss worse than hell

NIMRUTH
NIMRUTH
Joined: 24 Feb 05
Posts: 12
Credit: 1886
RAC: 0

okay... after several

okay...

after several crashes I have settled for 1710Mhz with a 190 FSB and a Vcore of 1.67 Volts..

I cranked it up to 1900 Mhz for a few minutes at 210fsb.. it crashed.. at 60C it shut down

I stand corrected for the Einstein WU time.. it takes me 7hrs36mins to complete.. But that`s at 1710mhz, once my watercooler is hooked up, I can run stable at 1850mhz or so, that would amount to aprox. 30mins speedup..

Basicly, all indications are that AMD beats the Intel P4 at any test..

I`ll never buy intel again..

Seti-Cruncher
Seti-Cruncher
Joined: 9 Feb 05
Posts: 70
Credit: 7114
RAC: 0

> Why is this? You would

Message 5784 in response to message 5782

> Why is this? You would think in Linux they could make it even better than for
> Windows.

Compiling for Linux is a lot more complicated than compiling for Windows because of the far more diverse range of platforms the compiler can target for. They simply haven't found the "right" compiler optimisations yet. Actually, it's even more complicated than that because the libraries linked (either dynamically or statically) need to be optimised too to have any real effect.

It's a shame the source code has not been released - otherwise some of the people out here with more time to spare may have cracked it by now. ;)

Be lucky,

Neil

Jordan Wilberding
Jordan Wilberding
Joined: 19 Feb 05
Posts: 162
Credit: 715454
RAC: 0

> Compiling for Linux is a

Message 5785 in response to message 5784


> Compiling for Linux is a lot more complicated than compiling for Windows
> because of the far more diverse range of platforms the compiler can target
> for. They simply haven't found the "right" compiler optimisations yet.
> Actually, it's even more complicated than that because the libraries linked
> (either dynamically or statically) need to be optimised too to have any real
> effect.

This still seems odd though. Because if the code is well-written, it should perform roughly the same as it does on windows, since both programs are basically spending most of their time doing Floating Point calculations, which is a CPU problem not a compiler problem.

I do agree they should release the source code. I think their arguments for not releasing it are kind of weak, so what if someone optimizes their client to get more points? Big deal, we are here to find gravity waves, not be the person with the most points!

such things just should not be writ so please destroy this if you wish to live 'tis better in ignorance to dwell than to go screaming into the abyss worse than hell

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.