Yes, I've never used optimised clients. I've also noticed very often the credit my machines claim is what everyone else gets.

Hi Sek, Electrolyte.

I feel that there is more of an issue here than most realise.
First, just to clarify a misconception, my experience has been, that it is the slow machines which claim the most points, not the fastest.

Also, and more to this issue, some machines seem to be penalised more than others. My prime example of this is a Celeron D360 2.4 Ghz machine that I run overclocked at 3.6Ghz. This machine has never ever recieved more points than it has claimed and very, very rarely gets its claimed points awarded. On rough calculations it is overall around 28%+ down on its claimed points over a considerable period of time.

I have other machines in a similar, but not as severe, situation.

Fact is the best reults I have are a break even senario on one machine of seven and the rest are all to some extent, averageing -15%, below claimed points.

The celeron has recently been switched to a different project, which does not rely on a quorum, and is producing at least twice the points that it was here. This is part of the optimization process that I am going through with the two BOINC projects which I support and another machine, which is more fairly treated by WCG has been brought to this project to replace the celery so don't panic.

The fact remains however that some computers and OSs are treated more and less fairly by this system. It stands to reason that if my computers are overall 15% down on claimed points and they are! Somebody else is basking in the reflected glow, so to speak.

I have put my hand up here to tell people what I am doing as far as project changing etc., is concerned and I leave myself open to nasty thoughts and comments as a consequence. Just keep in mind that I am not the only one who has done this type of thing and worse, but raise these issues so that they can be corrected before they cost the project too dearly.

Oh, of course the Rosetta project system is "very close to the '100,000 credit per Teraflop processed' principle." That's what it is designed to do!!

Cheers. ozylynx

Electrolyte, if u are saying that before u made 800-900 on Windows elsewhere and now 350 on Linux here, that would be the definitive answer to the massive drop......know people who switched back to WOS just for the points. From a crunching thruput, u should see the same, if not slightly better on Linux.

ciao

No, I wasn't using Windows before. ClimatePrediction and Einstein@Home ditched the benchmarks and use their own system. CPDN gives you credit depending on the position the WU (or model) is at and Einstein@Home give you the same amount of credit for WUs depending on their size - this also goes for Riesel Sieve.

Also, one of my team members has a dual-Xeon @ 3GHZ running WCG and on Windows, but he claims far more than me even though my FX-55 and P4 both perform better than his Xeon's. My P4 is basically a Xeon but for desktop computers (runs at 3.2GHZ and has 512KB L2, 2MB L3 cache on the Northwood core). The FX-55 is even faster than the P4.