Check out the new user who turned in 10000 results in 2 days from a 500MHz Pentium.
http://setiweb.ssl.berkeley.edu/show_host_detail.php?hostid=681542

> Check out the new user who turned in 10000 results in 2 days from a 500MHz
> Pentium.
> http://setiweb.ssl.berkeley.edu/show_host_detail.php?hostid=681542
>
>mmmmmmm Dont like the look off that

---

He didn't turn 'em in, he just got 'em.

he is one of the reasons, why 4 WUs are sent, although only 3 were needed for the quorum.

He has already crunched some, and has overcome former DL-errors with this flooding of his HD. (I tried by random his 1000-1020 results)

---

Gruesse vom Saenger

For questions about Boinc look in the BOINC-Wiki

Then expain to me why his computer is listed as the second most powerful in the program (see, http://setiweb.ssl.berkeley.edu/top_hosts.php)

> Then expain to me why his computer is listed as the second most powerful in
> the program (see, http://setiweb.ssl.berkeley.edu/top_hosts.php)
>
There has to be something wrong wit the data of this machine:

Owner Anonymous
Created 30 Mar 2005 0:00:33 UTC
Total Credit 214,359.66
Recent average credit 2,880.19
CPU type GenuineIntel 564MHz Pentium
Number of CPUs 1
Operating System Microsoft Windows XP Professional Edition, Service Pack 1, (05.01.2600.00)
Memory 126.42 MB
Cache 976.56 KB
Measured floating point speed 502.8 million ops/sec
Measured integer speed 1237.1 million ops/sec
Average upload rate 13.52 KB/sec
Average download rate 182.53 KB/sec
Average turnaround time 3.89 days
Results 10000

Perhaps he is a real veteran, just his joining date is mixed up, but I agree, the wizzards of Berkeley should take a look!

---

Gruesse vom Saenger

For questions about Boinc look in the BOINC-Wiki

Ok,

This system claims to take about 14:40 per WU...(This seems to be a Pentium II 500Mhz or Celeron Hydra)
And from a short count I saw 42 WUs returned on March 29th...but I stopped counting at 42.

Many of them had already been validated...

So, this represents a known issue...where a person can have many identical machines, and either
A. Copy the client_state.xml from their initially installed machine to all other machines in their farm
B. Install boinc regularly, attach to project on each machine, then on the boinc "my computers" web page, Merge these hosts into one host.

A hydra-host can download up to 100 WUs per day. So this would keep about 50 Pentium II 500Mhz class machines busy.

They are producing valid science, as the results are getting validated.
Their "machine" is probably higher in RAC than it should be.

There are others with this behaviour as well. Very strange.

http://setiweb.ssl.berkeley.edu/results.php?hostid=639525&offset=3680 to http://setiweb.ssl.berkeley.edu/results.php?hostid=639525&offset=4480 are all from 2 days (sent 12-13 March this year.
He claimed usually ridiculous small amounts of credit and got the usuall stuff.

> There are others with this behaviour as well. Very strange.
>
> http://setiweb.ssl.berkeley.edu/results.php?hostid=639525&offset=3680 to
...is valid. He really does have a 64 CPU Silicon Graphics, Inc. IP35 as far as I can determine.
His benchmarks suck thus low credit claims, but with 64 CPUs who would care.

Check out host #316624. Wow those are low CPU times for a P4 2.8Gig HT...hmm wait thats my host ;0

> > There are others with this behaviour as well. Very strange.
> >
> > http://setiweb.ssl.berkeley.edu/results.php?hostid=639525&offset=3680
> to
> ...is valid. He really does have a 64 CPU Silicon Graphics, Inc. IP35 as far
> as I can determine.
> His benchmarks suck thus low credit claims, but with 64 CPUs who would care.

OK :)
THX for the explanation :)
Blood pressure is down again ;)

@Benher

How on earth did yo umanage to get your 2.8's to get such high bench scores and compute so quickly??

Matt

---

> @Benher
>
> How on earth did yo umanage to get your 2.8's to get such high bench scores
> and compute so quickly??
>
> Matt

With the code I wrote and submitted to the seti@home project, which is now posted at setiboinc at sourceforge.

https://sourceforge.net/projects/setiboinc/
----
The benchmark scores are part of a change I made to BOINC, where each CPU regardless of speed will claim same ammount of credit for a WU.

(sematics: A given machine claims different ammounts for different WU times, of course, but if two of my machines happened to work on the same WU, they each would claim same credit for it. The average for all claimed credit is approx 32 for seti, ~20 for predictor.)