Hello Mark Reiss,
This has been asked for before. Implementing it would be a WCG-specific development project since no other group runs multiple projects on the same server. Our programming manpower is tied up on other projects. If we changed our scheduler this much, we would have to spend a lot of time reprogramming changes that BOINC made in the generic scheduler. I suppose I should say - more time than we do now. We already have a heavily customized scheduler.

WCG will do whatever it thinks will work best, but my personal guess is that it will only prioritize individual member's projects as part of a BOINC-wide development project rather than as a WCG-pecific development project.

Lawrence

Ingleside, WCG don't publish details of their customisations anywhere. From time to time the techs may offer you a glimpse, but that is the most you will get.

On the other hand, as responsible open source developers, they contribute back to BOINC any useful scheduler enhancements, such as the homogeneous redundancy feature.

As for the project ranking request: this is our most frequent FAQ where the answer is "not going to happen".

You can choose which projects you want to run. If you have lots of computers, you can weight projects by using different profiles.

But weighting in the scheduler would conflict with the other scheduler goals, and make an already difficult balancing act ridiculously difficult. It would also cause additional problems when members inevitably claim that their weightings are not being respected.

I shouldn't say this will never happen, but there are no plans for it, and it is not on any todo list.

Don't think anyone was asking about his/her ranking wishes, just a (repeated) suggestion for a method to order project work requests. Since another project with multiple was brought up, went to revisit and found no prioritization there either, but they do have too "Send work from any subproject if selected projects have no work"

At WCG, 9 active projects, could become 11, on 4 OS platforms, some having a homogeneous redundancy split down to processor families, think it is going to cause very major scheduling and load headaches.

Ok, then how about adding a few more profile slots? As it stands right now I have 6 computers doing assorted things around the house/work. Being able to set up a profile specific to each of them would allow the OCD tweaking that I want to be able to both work on the projects that I really want to & also to do some badge hunting.



Weighted scheduling (by percent of resources sent to each project) I agree would be a nightmare to make work. However it shouldn't be that difficult to implement a priority system of some sort.

(I'm not a programmer so you'll need to translate for those who are ) When setting up a profile the person would rank the projects 1-x. Using "if then" statements the program would go to #1 (say CEP) & if there was no work available then go to #2 (HCC) & repeat the process. Etc.

****

It's funny, over the years I have spent a lot of time either volunteering or sitting on boards that implement volunteer labor for some project (Scouting, church groups, library boards, etc). What's really sad is that almost always, on both sides of the table, there are hurt feelings because most can't see the forest for the trees. The volunteers forget that the end result is what they are working towards & that some of the things that they want to do really don't help reach that end. The professionals on the other hand, forget that the volunteers are just that - volunteers. If the organizers make things to difficult for them, eventually they will stop complaining & just walk away instead.

Sadly I've started to see the same thing here at WCG (perhaps it has been there for a long time but I've finally started seeing the forest). Because of the incredibly long term nature of all the projects, it is hard to see any short term progress. Because of that some of the crunchers (myself included ) like to fiddle/micro manage their settings so that they can actually see things relating to their input change. While the best thing may be to set up the computer once, turn it on & walk away to let it crank untouched for a few years - that's probably not going to happen.

On the tech side of things, the attitude of "we're saving the world, don't bother us" isn't going to do much towards keeping people willing to help. If it's going to take years if not decades to make progress on the problems that all this computer time is being devoted to, then spending/investing towards good will a week or two of programming time to give the volunteers shiny buttons & knobs to push & turn to their heart's content is at the very least a smart thing to do (it might even be considered an essential part of the process).

Anyway, just my 2% of a $ for the morning...