Message boards : Number crunching : Overclocking?....Not! ;)
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile rbpeake

Send message
Joined: 17 Sep 04
Posts: 99
Credit: 30,836,799
RAC: 9,863
Message 18154 - Posted: 14 Oct 2007, 18:40:13 UTC
Last modified: 14 Oct 2007, 18:41:18 UTC

It would appear from my abysmal record on this machine which, while overclocked, does very well on Rosetta, Einstein, and many other projects, is not appropriate for this project! Even one of the "pending" units waits in limbo while all the others have already received credit! ;)
Regards,
Bob P.
ID: 18154 · Report as offensive     Reply Quote
KAMasud

Send message
Joined: 7 Oct 06
Posts: 114
Credit: 23,192
RAC: 0
Message 18156 - Posted: 14 Oct 2007, 19:10:58 UTC


:-) Lucky you ;-) mine used to give client error when i was over clocking.
Regards
Masud.
ID: 18156 · Report as offensive     Reply Quote
Toby

Send message
Joined: 1 Sep 04
Posts: 137
Credit: 1,691,526
RAC: 10
Message 18161 - Posted: 15 Oct 2007, 4:39:25 UTC

I believe LHC is much more sensitive to floating point errors than other projects. They have even had problems with the slight differences between the way Intel and AMD implement floating point math. Your overclocked CPU is probably returning slightly incorrect results on other projects as well but they match closely enough that you still get credit. The thing with LHC is that if the first calculation is off by even 0.0000[...]001 then by the end you are way off because later calculations are all based on previously calculated numbers.
- A member of The Knights Who Say NI!
My BOINC stats site
ID: 18161 · Report as offensive     Reply Quote
Brian Silvers

Send message
Joined: 3 Jan 07
Posts: 124
Credit: 7,065
RAC: 0
Message 18162 - Posted: 15 Oct 2007, 7:39:28 UTC - in response to Message 18161.  

I believe LHC is much more sensitive to floating point errors than other projects. They have even had problems with the slight differences between the way Intel and AMD implement floating point math. Your overclocked CPU is probably returning slightly incorrect results on other projects as well but they match closely enough that you still get credit. The thing with LHC is that if the first calculation is off by even 0.0000[...]001 then by the end you are way off because later calculations are all based on previously calculated numbers.


I too have an overclocked system and am getting a 50% invalid ratio at this point with LHC. I don't think that this happened before the move to the UK, but yes, I realize that with extended time of overclock/overvolt, problems can begin to happen...

Question to Alex/Neasan:

Are the validators the same as what they were before the move, or has there been any change made to the validator code?
ID: 18162 · Report as offensive     Reply Quote
zombie67 [MM]
Avatar

Send message
Joined: 24 Nov 06
Posts: 76
Credit: 7,919,897
RAC: 789
Message 18165 - Posted: 15 Oct 2007, 14:17:55 UTC

I have two OC machines attached, and they are doing just fine. No validation problems at all.
Dublin, California
Team: SETI.USA

ID: 18165 · Report as offensive     Reply Quote
daemion

Send message
Joined: 4 Dec 06
Posts: 33
Credit: 75,491
RAC: 0
Message 18166 - Posted: 15 Oct 2007, 15:47:03 UTC

My Athlon 2800, which is OCed to about 3600, does fine on every project i have it working for. although i have noticed that if i try to play a game, when the cpu is currently crunching a LHC WU it inevitably causes it to error out. LHC is the only project that this happens for. so it seems that LHC WUs are a great stability test for your CPU. i had mine up to 3800, crunched away nicely until it got LHC WUs then it errored all of them. i dropped the clock and everything is back to normal, except when i play games of course :)
ID: 18166 · Report as offensive     Reply Quote
Profile Neasan
Volunteer moderator
Volunteer tester
Avatar

Send message
Joined: 30 Nov 06
Posts: 234
Credit: 11,078
RAC: 0
Message 18167 - Posted: 15 Oct 2007, 16:38:37 UTC

It is the same validator as before the move. As mentioned LHC@home is uber-sensitive to slightly mismatched results so over clocking can affect validation for some people it works and some it doesn't this was the case before the move and is still the case now.
ID: 18167 · Report as offensive     Reply Quote
Bryan

Send message
Joined: 17 Nov 07
Posts: 2
Credit: 169
RAC: 0
Message 18685 - Posted: 10 Dec 2007, 12:45:31 UTC

I have just started using LHC but on World Community Grid if my cpu temp goes over 130F then I start getting errors on about 1 out of 4 results. Better cooling may sovle your problem.
ID: 18685 · Report as offensive     Reply Quote
Profile meckano
Avatar

Send message
Joined: 17 Sep 04
Posts: 150
Credit: 20,315
RAC: 0
Message 18796 - Posted: 2 Jan 2008, 21:52:46 UTC

depends on definition of overclock...

I've never own'ed an AMD 3200+ cpu, it has always been an:
AMD 2500+ 333mhz.
I got lucky with an underclocked model, and with 2 jumpers in the cpu socket,
I got it to run at 400mhz bus speed, and because the mutiplier is the same as
that of the 3200+, all is good. (not even a temperature increase :D )
I did much checking when I started, every time I had a glitch, but it always
turned out to be a software issue; so it remains um, boosted.
I did nothing else, and not a program yet has stated it to be anything other
than a 3200+. (was glad I bought all 400mhz ram, yet the system could run
the ram at 333 if need be.)

While I'm on odd things, the Kingston ram for this system CL3 does not work;
maybe due to the Corsair being CL2.5 @ 400mhz...
... So I was given a Kingston CL2.5 to try, even though my mobo ain't on its
list, and it works.
I now have 3 x 512, 1 Corsair, 2 Kingston, at CL2.5 @ 400mhz with 2.2 ghz cpu.

-----------------------
Click to see my tag
My tag
SNAFU'ed? Turn the Page! :D
ID: 18796 · Report as offensive     Reply Quote

Message boards : Number crunching : Overclocking?....Not! ;)


©2024 CERN