Check out this 10 Gig Machine
http://setiathome.berkeley.edu/show_host_detail.php?hostid=1735123

---

I had a quick glance and it looks like he processes WUs a tad slower than my P4 2.6.

But maybe I overlooked sumthin.

---

mambo

CPU type GenuineIntel
Intel(R) Pentium(R) 4 CPU 2.80GHz
Number of CPUs 2
Operating System Microsoft Windows XP
Professional Edition, Dodatek Service Pack 2, (05.01.2600.00)
Memory 502.79 MB


It says it is only a Pentium 4 2.8GHZ Machine Hyperthreaded with 512Megs of Memory. He is taking 13,000 Seconds per Work Unit which is on par for a Non Optimized Machine. Optimized he could lower that to the 6,500 to 7,500 Second range.

---

I'd say that's a bad or 'adjusted' ;) benchark score...

Most of his claimed credits look reasonable (except 1 or 2) so maybe something that will correct itsef next time the machines benchmarks run...

cheers,

Paul.

---

Wanna visit BOINC Synergy? Click my stats!

Join BOINC Synergy Team

He is just adjusting his benchmark score to claim "more" credit. I would love to see the look of surprise on his face when he relises he's not getting the credit he claimed. LOL.

---

hmm... perhaps its for the beniit of another project while SETI is down.. one that has a quorum = 1 and you 'gets whatchya claim'....

I would like to think not...

Paul

---

Wanna visit BOINC Synergy? Click my stats!

Join BOINC Synergy Team

CPU type GenuineIntel
Intel(R) Pentium(R) 4 CPU 2.80GHz
Number of CPUs 2
Operating System Microsoft Windows XP
Professional Edition, Dodatek Service Pack 2, (05.01.2600.00)
Memory 502.79 MB


It says it is only a Pentium 4 2.8GHZ Machine Hyperthreaded with 512Megs of Memory. He is taking 13,000 Seconds per Work Unit which is on par for a Non Optimized Machine. Optimized he could lower that to the 6,500 to 7,500 Second range.

Out of interest. how would you optimise the machine?

---

Richard King

Notts PC Services

Optimise: download a client dedicated to SSE2 Pentium 4 processor =)

He has got good benchmarks, but he's WU completing times are 6-7 times greater than on my A64 :D

Those benchmarks wont give him anything... he got claimed 230 but will gain 30 average at most...

---

From what I see looking at this machine, its benchmark scores, and its claimed credit per result, with quite a few claiming more than 400 credits per result, it would seem that he is running an optimized BOINC client that has corrupted the benchmark routines during the compile, or that the client source code was altered intentionally and recompiled to artificially inflate the claimed credit. I don't want to make any accusations though, as I don't have all the facts.

This is a prime example for those who question why the validation process tosses the highest and lowest claimed credit and takes the middle value if 3 results form the quorum, or averages the two middle values if 4 results form the quorum.

Regards, Daniel.

---

From what I see looking at this machine, its benchmark scores, and its claimed credit per result, with quite a few claiming more than 400 credits per result, it would seem that he is running an optimized BOINC client that has corrupted the benchmark routines during the compile, or that the client source code was altered intentionally and recompiled to artificially inflate the claimed credit. I don't want to make any accusations though, as I don't have all the facts.

This is a prime example for those who question why the validation process tosses the highest and lowest claimed credit and takes the middle value if 3 results form the quorum, or averages the two middle values if 4 results form the quorum.

Regards, Daniel.

looking at this machine, I would say he is running a default client and a default app, but ran the benchmark with hyperthreading disabled and then manualy shifted the decimal point in the benchmark-scores 1 position.

From what I see looking at this machine, its benchmark scores, and its claimed credit per result, with quite a few claiming more than 400 credits per result, it would seem that he is running an optimized BOINC client that has corrupted the benchmark routines during the compile, or that the client source code was altered intentionally and recompiled to artificially inflate the claimed credit. I don't want to make any accusations though, as I don't have all the facts.

This is a prime example for those who question why the validation process tosses the highest and lowest claimed credit and takes the middle value if 3 results form the quorum, or averages the two middle values if 4 results form the quorum.

Regards, Daniel.

looking at this machine, I would say he is running a default client and a default app, but ran the benchmark with hyperthreading disabled and then manualy shifted the decimal point in the benchmark-scores 1 position.

Sounds logical. These are the people that should be remove from the project for "cheating".

---

I have a pentium 4 running at 3.06Ghz and it does work units in about 2hrs, my freind has a 1.5Ghz Celeron M doing work units in the same time? I figured optimisation had something to answer for. If you know of one could you E-mail a link for a pentium 4 Seti@home optimiser please to m.k.bowler@hotmail.co.uk many thanks...

---

Readin all this i guess i optimized my system very well,
getting an average credit around 20 for each work unit.
Not much, but really running fast through all the units within an average of 40min to calculate.

---

All that we see or seem is but a dream within a dream.