Hi, I'm trying to understand work done, to gauge whether it's worth keeping certain computers powered up. I have questions...

The manager lists "Work done" for me as 268,571.97 is that measured in Cobblestones?

The documentation refers to "Recent Average Credit" but the manager has a column for "Avg. work done". Is that the same thing? If it's an average, what is it averaged over? Is it Cobblestones per day, or Cobblestones per job, or what?

I was going to compare the efficacy of my different computers by seeing how long each job takes; however, this calculation will be screwed if the job size is scaled to the compute power. Are all jobs the same, or are they chosen with the computer in mind?

The Rosetta project says that the total credit is 20,112,777,437 and my Work Done is 268,571.97. Does that mean that I've achieved 0.001335% of the total work, or a 1/74,888th part?

Thanks

The manager lists "Work done" for me as 268,571.97 is that measured in Cobblestones?

Yes, Cobblestones or just Credit. They're both the same, really.

The documentation refers to "Recent Average Credit" but the manager has a column for "Avg. work done". Is that the same thing? If it's an average, what is it averaged over? Is it Cobblestones per day, or Cobblestones per job, or what?

RAC, Recent Average Credit and Average Work Done are all the same, but for different BOINC versions. The developers change the wording just about every major release, and it gets difficult keeping up with all the changes in the manual.

For most of the answers, see http://boinc.berkeley.edu/wiki/Computation_credit

Are all jobs the same, or are they chosen with the computer in mind?

You can't compare work between projects. You can't even compare all work done within one project, as a lot of projects use different applications, doing different work.

To measure how well your computer is doing, the best way is to check its host RAC. A high one means your computer is doing quite well. But that depends on which project you run as well, plus how many GPUs you have these days.

If you run e.g. Einstein BRP4s on the GPU, these tasks take mere minutes these days on a good GPU. And each of them pays 500 credit when validated. If you then run 24/7, with one or more BRP4s running per GPU, you get a very high RAC in no time (two weeks to max out, still).

But you can't compare the BRP4s with the SL6Vs that Einstein also runs. Or the Astropulse tasks at Seti. Or any of the applications at WCG. It's something that the developers are trying to make a system for, that eventually all projects will have to go use, but in the mean time, while it's not there yet, there's no really good way to measure efficiency between projects.

---

Interesting answer; thanks. Yes, it's important when explaining things, to be consistent with names, because for novices, it's like a jigsaw puzzle, and the pieces need to actually fit to get the picture. Especially for people who are peripherally interested, not enthusiasts.

But you can't compare the BRP4s with the SL6Vs that Einstein also runs. Or the Astropulse tasks at Seti. Or any of the applications at WCG. It's something that the developers are trying to make a system for, that eventually all projects will have to go use, but in the mean time, while it's not there yet, there's no really good way to measure efficiency between projects.

I hope they cut at least two or three decimal points off the score while they are at it. The speeds of GPUs (not to mention the advances in CPUs) make the point scores ridiculous, and it is hard to compare them.