S40 Observation Thread

Erik
Erik
Joined: 14 Feb 06
Posts: 2,815
Credit: 2,645,600
RAC: 0

RE: AGI (address generation

Message 27316 in response to message 27315

Quote:
AGI (address generation interlock) isn't depend on the type of motherboard.
It comes from unlucky instruction order and register usage.

Ok, thanks. The only AGI I knew of was Asrock's variant of Acc. Graphics Port. I'd read some posts about using graphic cards for processing.

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4,527,270
RAC: 0

RE: Hello! S40 did some

Message 27317 in response to message 27306

Quote:
Hello! S40 did some "access violation" error on one of my PC.
So, don't use this version! I will check the code and the PC.

I did a fast code checking. It seems to be good. :-(

LiborA
LiborA
Joined: 8 Dec 05
Posts: 74
Credit: 337,135
RAC: 0

RE: RE: Have somebody got

Message 27318 in response to message 27313

Quote:
Quote:

Have somebody got access violation error with S40?/quote]

No problem here with 24 WU's completed and only a couple of those pending. Still running about 20 minutes faster than S39L. These WU's are in numerical order so they are all about the same length.

A have not problem also. On my Xeon S40 has uniform speed as S39L
S40: avg. 6030,71 sec (avg. from 11 WUs)
S39L: avg. 6034,55 sec (avg. from 19 WUs)

Misfit
Misfit
Joined: 11 Feb 05
Posts: 470
Credit: 100,000
RAC: 0

I didn't get any errors with

I didn't get any errors with it but it was around 100 seconds slower than S39L on my Prescott (HT enabled).

me-[at]-rescam.org

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4,527,270
RAC: 0

RE: I didn't get any errors

Message 27320 in response to message 27319

Quote:
I didn't get any errors with it but it was around 100 seconds slower than S39L on my Prescott (HT enabled).

Thanks.
S40 is also slower on my P4-M Northwood (~1%).

The prefetch logic / cache mechanism of AMD processors is much better than intel.
Probably, intel forgot something. (AMD -> write combining)

Raistmer*
Raistmer*
Joined: 20 Feb 05
Posts: 208
Credit: 180,190,059
RAC: 11,767

2WUs done on 2 different

2WUs done on 2 different processors (Thoroughbred, Winchester), 1 verified - no problems so far.

Dave Burbank
Dave Burbank
Joined: 30 Jan 06
Posts: 275
Credit: 1,548,376
RAC: 0

Has anyone who is running the

Has anyone who is running the D40 application tried the S40? And if so how do they compare?

Thanks

There are 10^11 stars in the galaxy. That used to be a huge number. But it's only a hundred billion. It's less than the national deficit! We used to call them astronomical numbers. Now we should call them economical numbers. - Richard Feynman

Raistmer*
Raistmer*
Joined: 20 Feb 05
Posts: 208
Credit: 180,190,059
RAC: 11,767

RE: Has anyone who is

Message 27323 in response to message 27322

Quote:

Has anyone who is running the D40 application tried the S40? And if so how do they compare?

Thanks


Differences are still in error range.

And some "bug" - after uploading calculated WU no new WU calculation was started by S-40 until I noticed it few hours later and pressed "Update project" button. No errors in messages were shown. There is no such bug on secondary PC.

Akos Fekete
Akos Fekete
Joined: 13 Nov 05
Posts: 561
Credit: 4,527,270
RAC: 0

So, probably I found the

So, probably I found the reason of "access violation" fault. There were two chances. Software or hardware fault.

S40 is running on my five computers and I got 6 faults from 200 wus.
I got the faults one the same computer, so it gives much bigger probability to hw fault.

But, I got the same fault. So I examined the code and the CPU more watchfully.
I found that the AGU part of the CPU cannot tolerate the higher frequencies.

Raistmer*
Raistmer*
Joined: 20 Feb 05
Posts: 208
Credit: 180,190,059
RAC: 11,767

Toroughbred 2400+ D-40 5010

Toroughbred 2400+
D-40 5010 (10 WU av)
S-40 4843 (2 WU av) (all 2 WU faster than all 10 of D-40)
It seems S-40 faster :)

Winchester on 213 HT
S-39L 4034 (12 WU av, very small WU time difference)
S-40 3854 (5 WU av)
S-40 faster here too. And no faults on slightly overclocked Athlon64 still.

Comment viewing options

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