I think that rutime is the best indicator of personal efficiency.

1.) Once you get the OS tweaked, it will likely stay the same until a major SP or entire OS upgrade. Can be eliminated as a factor in day to day comparisons.

2.) Once you get your hardware tweaked (quit playing with the OC and just let it crunch) that will not likely not change often. Can be eliminated as a factor in day to day conparisons.

3.) Once you determine which project(s) run best on your machine that will likely not change until you change #1 or #2.

Once a system is settled on all the aboove points and you run the same project(s) day after day, your points will settle in also.

At that point, given a stable environment, the only thing you can do to effect how much science you are processing is how many hours a day you let it run. You can see stats for each computer in the Miscellaneous section of your "My Grid", "Statistics", click on "Device Installations", and finally click on your individual computer names.

The last thing to factor in is that these numbers come from when your WUs are validated. Unless you manually track every WU you process you will always have a little variation that you can do nothing to control. That being said, if you leave everything alone for 1 week, you should start to get some very stable stats.

If you are not running at 95%+ you need to leave it turned on longer
Just kidding, these machines are ours to use and the crunching is what happens in between

To me production is all that matters and assuming that no matter what project you do they are awarded the same( they aren't but we have to make some assumptions) then if you want to see how your machine does compared to others then you look here and use the credit/month tab as that esentially minimises any variances due to people storing completed WU's and uploading say a weeks worth in one day.
http://boincstats.com/stats/host_stats.php?pr...active=&st=0&or=8

Movieman for an absolute comparison I agree that boincstats is ok. The question here is to have a relative measure about the maximum capability of your own rig, and not another similar. Every rig is different even if you have the same processor: HT or not, memory quantity and speed, OC, Disk type and speed, motherboard, network, OS configuration, GPU etc. you name it.
What I am trying to do is to have a personal reference in terms of possible maximum and then compare daily or weekly how the whole system behaves, including whatever interruptions maybe. This is a moving target and the maximum is also depending on the project, or if you run a single or mixed projects. But over a month of stable usage by taking out the best day of each you get that relative max, you can compare the other days.

Snow Crash I think that in terms of reliability points and runtime are similar. If I am not wrong and then please correct me, but what you are awarded in terms of points and runtime is based on VALID returned results. This is depending on many external factors related to WCG (waiting for the duplicates to finish, quorum, connection frequency to report results etc.). Even if on a continuous basis it flattens all these issues still they affect both points and runtime.
I have 8 machines that run 24/7 (and another two that are also used for personal work so they are less reliable) and even when they run without any interruption at all I have variations in runtime.
As a recent example my MARS, TERRA, MERCURY devices (i7 950) have run 24 hours exclusively on FAAH this Monday and Tuesday without any interruption or external influence of any kind:

MARS
Monday - runtime 0:008:16:12:02
Tuesday - runtime 0:007:19:55:00
approx 1 day difference

TERRA
Monday - runtime 0:007:06:26:38
Tuesday - runtime 0:009:19:55:00
approx 2 days difference

MERCURY
Monday - runtime 0:007:12:20:21
Tuesday - runtime 0:008:16:41:00
approx 1 day difference


Same for SATURN (i7 950) but here the numbers are very near to each others.

SATURN
Monday - runtime 0:007:07:15:56
Tuesday - runtime 0:007:07:56:51

Here there is only 41 minutes difference which is remarkable.

Sek I did exactly what you said.
But when I click the Host List, it is an empty page I get. There are no hosts at all. Ther must be something else to do.

Thnk's for the reply Sek. I have in BAM created an account with WCG as the unique project. But still this does not change anything.
I suppose you must attach the hosts one by one. There must be a way to attach all hosts which are on a WCG account in one shot. Maybe hiding issues in WCG affect that.
The way to attach the hosts is not straightforward at all and the FAQ in BAM is also not very clear. I have to probe further and invest more time ,maybe this coming week-end.