Weird benchmark.

Message boards : Number crunching : Weird benchmark.
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Paul Knibbs

Send message
Joined: 21 Mar 01
Posts: 5
Credit: 194,839
RAC: 0
United Kingdom
Message 51789 - Posted: 6 Dec 2004, 22:07:00 UTC

Why is the computer I just installed BOINC on showing a dhrystone number of over 113,000? I've run the benchmark on it twice now and it comes out the same each time! No way a Sempron 2400+ (or, in fact, any computer currently existing on the face of the planet!) should be able to return those sort of numbers for a single CPU...
ID: 51789 · Report as offensive
Profile littleBouncer
Volunteer tester
Avatar

Send message
Joined: 28 May 99
Posts: 151
Credit: 666,283
RAC: 0
Switzerland
Message 51790 - Posted: 6 Dec 2004, 22:09:07 UTC - in response to Message 51789.  
Last modified: 6 Dec 2004, 22:21:01 UTC

> Why is the computer I just installed BOINC on showing a dhrystone number of
> over 113,000? I've run the benchmark on it twice now and it comes out the same
> each time! No way a Sempron 2400+ (or, in fact, any computer currently
> existing on the face of the planet!) should be able to return those sort of
> numbers for a single CPU...
>
>

That's a beauty can U send it to me ?:-)
(I mean the CPU)

[EDIT] That is a joke, sorry.

ID: 51790 · Report as offensive
Profile Paul Knibbs

Send message
Joined: 21 Mar 01
Posts: 5
Credit: 194,839
RAC: 0
United Kingdom
Message 51793 - Posted: 6 Dec 2004, 22:17:48 UTC

Well, thanks for that, but I'm kind of worried--this machine is going to be reporting some ludicrously high "claimed credits" and they might all be disallowed, or something!
ID: 51793 · Report as offensive
SURVEYOR
Volunteer tester

Send message
Joined: 19 Oct 02
Posts: 375
Credit: 608,422
RAC: 0
United States
Message 51817 - Posted: 7 Dec 2004, 2:20:52 UTC

better set connect every 0.01 days or will down load all of the wu's
ok there is suppost to be 50 file limit.
Fred
BOINC Alpha, BOINC Beta, LHC Alpha, Einstein Alpha
ID: 51817 · Report as offensive
Divide Overflow
Volunteer tester
Avatar

Send message
Joined: 3 Apr 99
Posts: 365
Credit: 131,684
RAC: 0
United States
Message 51819 - Posted: 7 Dec 2004, 2:24:47 UTC

Are you using one of the custom optimized linux BOINC clients out there? Some of these optimizations are actually breaking the BOINC benchmark code. The benchmark delibrately asks the computer to run some non-optimized code that custom complied jobs can possibly re-write, giving inflated benchmark results.

Crunch a few WU's and see how it does!

ID: 51819 · Report as offensive
Profile Paul Knibbs

Send message
Joined: 21 Mar 01
Posts: 5
Credit: 194,839
RAC: 0
United Kingdom
Message 51937 - Posted: 7 Dec 2004, 7:10:43 UTC

It's the Linux client downloaded from this very site, and it looks like whatever timing glitch causes the benchmark to go wrong is also affecting the timer for the actual crunching, so they cancel each other out--it's so far crunched 2 units, and it claims they each took around 270 seconds to crunch, even though I know for a fact they must have taken several hours (the machine's been running all night!). Claimed credit is around 25 per unit, which is a little lower than my main crunching machine but not by a massive amount.
ID: 51937 · Report as offensive
Profile Paul Knibbs

Send message
Joined: 21 Mar 01
Posts: 5
Credit: 194,839
RAC: 0
United Kingdom
Message 51939 - Posted: 7 Dec 2004, 7:36:53 UTC

Update: I just applied some updates to the Red Hat Fedora installation I have, and the problem seems to have reversed itself--I'm now showing just 2000 Dhrystones, whereas a 2400+ should be good for nearer 3500 given some other machines I've seen!
ID: 51939 · Report as offensive

Message boards : Number crunching : Weird benchmark.


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.