S39L Observation Thread

Raistmer*
Raistmer*
Joined: 20 Feb 05
Posts: 208
Credit: 179937384
RAC: 63097

RE: The link is live again.

Message 26433 in response to message 26430

Quote:
The link is live again. (SETI K7-08)
But I don't do any funny remarks and I don't promise anything.
Is okay?


Sure :) Thank you.

Quote:
did a few WUs with that app. They finished OK but took 7-10 min longer.


Ok, my own statistic:

ClawHammer
Crunch3r SSE2 3692 (av 12)
akosf K7-08 3447 (av 32)

Winchester
Crunch3r SSE2 2955 (469 av)
akosf K7-08 2865 (19 av) after little overclock (2040MHz->2130MHz) 2653 (78 av)

Venice
Crunch3r SSE3 3001 (95 av)
akosf K7-08 2780 (47 av)

Prescott
Crunch3r SSE3 3244 (53 av)
akosf K7-08 3201 (76 av)

I dont like to stop calculations for testing reference WU so only averaged values available. All noisy very short WUs were removed.

So maybe on some processors improvement is small but on all procs it's improvement in speed.
Let K7-08 be alive :)

Raistmer*
Raistmer*
Joined: 20 Feb 05
Posts: 208
Credit: 179937384
RAC: 63097

RE: Prescott Crunch3r SSE3

Message 26434 in response to message 26433

Quote:
Prescott
Crunch3r SSE3 3244 (53 av)
akosf K7-08 3201 (76 av)


Crunch3r SSE3 3368 +/- 59 (51 av) - 2 bad WU were included, sorry.
akosf K7-08 3201 +/- 53 (76 av)

So on P4 improvement is noticed too.

Misfit
Misfit
Joined: 11 Feb 05
Posts: 470
Credit: 100000
RAC: 0

akosf, which has greater

akosf, which has greater accuracy. S39L or S40.03?

me-[at]-rescam.org

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4527270
RAC: 0

RE: akosf, which has

Message 26436 in response to message 26435

Quote:
akosf, which has greater accuracy. S39L or S40.03?

S40.03 has fewer multiplication, so it means fewer rounding.
But this difference is very very small.
It means that each 2048th calculated result has one bit difference. :-)

Misfit
Misfit
Joined: 11 Feb 05
Posts: 470
Credit: 100000
RAC: 0

RE: RE: akosf, which has

Message 26437 in response to message 26436

Quote:
Quote:
akosf, which has greater accuracy. S39L or S40.03?
S40.03 has fewer multiplication, so it means fewer rounding.
But this difference is very very small.
It means that each 2048th calculated result has one bit difference. :-)


And how about the cache requirements? Does S40.03 use a smaller amount of cache like S39L?

me-[at]-rescam.org

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4527270
RAC: 0

RE: And how about the cache

Message 26438 in response to message 26437

Quote:
And how about the cache requirements? Does S40.03 use a smaller amount of cache like S39L?

Both versions use same algorithms and same amount of data.

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4527270
RAC: 0

RE: Better investigate on

Message 26439 in response to message 26318

Quote:
Better investigate on Linux, the growing distance is really frustrating, on the other hand I'm happy to "beat" these 4-cylinder Macs with your albert and my X2. :))

I just see, your twin-cylinder Athlon is flyable. :)

M. Schmitt
M. Schmitt
Joined: 27 Jun 05
Posts: 478
Credit: 15872262
RAC: 0

RE: RE: Better

Message 26440 in response to message 26439

Quote:
Quote:
Better investigate on Linux, the growing distance is really frustrating, on the other hand I'm happy to "beat" these 4-cylinder Macs with your albert and my X2. :))
I just see, your twin-cylinder Athlon is flyable. :)

:-)

Yes, he got a good cooler now(Zalman 9500) and runs at 2700MHz without problems. :)
Isn't it a good thing to run a cpu at the speed of a twice as expensive model?
I know, you like to do so, too. ;)

Did you hear that in the second half of the year the new Athlon K8L(65nm SOI) is coming, "Pacific" included? Then my X2 is an old iron again. ;)

Last not least I want to thank you for the new turboalbert S40.04.

Relaxing holidays!
Michael

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4527270
RAC: 0

RE: Yes, he got a good

Message 26441 in response to message 26440

Quote:
Yes, he got a good cooler now(Zalman 9500) and runs at 2700MHz without problems. :)
Isn't it a good thing to run a cpu at the speed of a twice as expensive model?
I know, you like to do so, too. ;)

You should try out a water (or a compressor) cooler. A simple water cooler isn't to expensive (60-70EUR), but the effect looks like an afterburner...

Quote:
Did you hear that in the second half of the year the new Athlon K8L(65nm SOI) is coming, "Pacific" included?

Of course, but Conroe will win, much.

M. Schmitt
M. Schmitt
Joined: 27 Jun 05
Posts: 478
Credit: 15872262
RAC: 0

RE: RE: Yes, he got a

Message 26442 in response to message 26441

Quote:
Quote:
Yes, he got a good cooler now(Zalman 9500) and runs at 2700MHz without problems. :)
Isn't it a good thing to run a cpu at the speed of a twice as expensive model?
I know, you like to do so, too. ;)
You should try out a water (or a compressor) cooler. A simple water cooler isn't to expensive (60-70EUR), but the effect looks like an afterburner...

I don't like the extra power consumtion. And upgrading the system to, for example a two socket board for dual/quad core opterons, would be more complicated. This way I could sell board, CPU, cooler and memory rather easily and upgrade.

Quote:
Quote:
Did you hear that in the second half of the year the new Athlon K8L(65nm SOI) is coming, "Pacific" included?
Of course, but Conroe will win, much.

65nm SOI is a much better technology than plain 65nm. ;)
So I'm not shure if you're right. Let's wait and get surprised. :)

cu,
Michael

Comment viewing options

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