FYI:
My win boxe (real dual core) does get 20.
My Linux boxes, all true SMP boxes are lucky to get one or two WU's per request.

Is there any cure for this?

Other factors that can drive this is availability, resource share, and other projects on the system.

---

Travis, any further word on when we might go to 20 work units per core or at least a reduced timeout before we can get more work?

Our quad core and better machines could do a lot more work for this project if at least the forced timeout between connections is reduced to something like 5 or 10 minutes.

---

Join the #1 Aussie Alliance on MilkyWay!

[quote]

when we update the server code, you should be able to download new workunits as soon as you finish with your previous ones. if you're just complaining about wanting more WUs out there in the case when the server crashes... i don't think theres anything we can do about that as the workunits need to be dynamically generated as new ones come in.

Are you going to set it up like burp ?

I have always felt that there should be a top limit to downloads of wu's on all BIONC sites.

Looking at bioniclogX complete times of 14:??:??. Could you just imagine the caos of haveing downloaded 700 wu;s as some hosts seam to be able to over on seti@home and then just sit on them. Even with a Q6600 I often have to NNT just to clear other Work. My activity is always run always connected and.1000 conect time. At this most times there are 20 Wu's crunching or waiting to run.

Travis, any further word on when we might go to 20 work units per core or at least a reduced timeout before we can get more work?

Our quad core and better machines could do a lot more work for this project if at least the forced timeout between connections is reduced to something like 5 or 10 minutes.

Me too as I have 4 Quads and one is more hungry than the others(slightly now, due to heat problems).

---

What ever happened to the 20 wu limit?

I've just received 40 wus at last contact with server?

---

What ever happened to the 20 wu limit?

I've just received 40 wus at last contact with server?

It's now 20 per core not per host.

arhhhhhhhh

should have read this post a bit better

---

Oh darn. No B Studio for Vista.

---

me@rescam.org

Right, my dual core, tri-core and quad core processors sucked up a lot more work units as a result of the change. Now if I had milsops custom application, I'd be generating Carl Sagan like numbers <smile>.

It's now 20 per core not per host.

---

I'd be generating Carl Sagan like numbers <smile>.

Billions and billions and billions.

---

me@rescam.org

Maybe the 20 workunit is now too much. I joined this project some time ago, and left it. In the meantime I forgot why and I reattached. Now I remember why I left.

On all my Pjuters Milkyway is running all the time in High Priority mode because my Pjuters download too many tasks.I don't have the time, and neither do I wish to, micromanage the Boinc projects I am supporting. I do know that when one project is constantly running in HP mode other projects lose dedicated time. When I receive about 21 tasks, with a deadline of about 48 hours (or 24 or so) and with an average computation time of about 7 hours, then when all my pc's have a dual processor this project is running HP all the time on all Pjuters.

This project deserves in my opinion a certain percentage of my time. Not all of it. Is it possible to either decrease the number of workunits, or to have your server make basic calculations about estimated finishing time of the tasks and stop forcing Pjuters getting HP's from Milkyway?

Again, I do not have time to micromanage, and neither do I have the time to follow up on the message boards. If I get an answer here I'm not certain if I will read it within 14 days from now. What I do know is that I will leave this project if, when I'm back at my office some 14 days from now, Milkyway is still running in HP on my Pjuters.

I appreciate every result being meaningful. I do not appreciate having a project "stealing" time from other projects by forcing High Priority runs all the time. And even less getting tasks that will time out anyway.

My 2 €urocents...

Maybe the 20 workunit is now too much.
[..]
When I receive about 21 tasks, with a deadline of about 48 hours (or 24 or so) and with an average computation time of about 7 hours, then when all my pc's have a dual processor this project is running HP all the time on all Pjuters.

I would wait with that judgement until you tried the new app. With that, the computation times are back to some minutes and 20WUs aren't that much after all.

Willing to wait and see. I do not mind getting 2, 20 or 200 tasks of 2, 20 or 200 hours each. I do mind a project cheating itself into more time by High Prioritizing its tasks.

Anyway: waiting...

I do mind a project cheating itself into more time by High Prioritizing its tasks.

Anyway: waiting...

While you're waiting, like many of us, perhaps you could read up on BOINC and Short Term / Long Term Debt. What actually happens if projects go into High Priority is that your other projects will actually get more time after the higher priority work has been cleared. The percentages you allocate are indeed honored, just not on a minute-by-minute timeframe. It might actually be week-by-week or month-by-month, but if left alone, your percentages do end up being honored...

Brian

When there was work, and the recent fast client was released, the 20 WU limit was very much to be noticed. It was a good working limit that prevented people being greedy and drawing down a large cache of WUs which the client was fully capable of completing by the deadline. This limit was a good thing and helped others get work.

I am running several projects at the same time, similar to rembertw. Occasionally the high priority was used by BOINC Manager, and my other projects never seemed to loose out (as far as I was aware).

I don't mind only getting 20 tasks, but I do mind 20 tasks that take 30 hours each. Especially when they are all due in 3 days. The optimized apps run much nicer, but there isn't enough work for them.

Have the Admins set a new lower WU-per-CPU limit?

I have just gotten this script from my BOINC Manager /Messages/ tab -

29/11/2008 21:17:55|Milkyway@home|Deferring communication for 35 min 8 sec
29/11/2008 21:17:55|Milkyway@home|Reason: no work from project
29/11/2008 21:18:01|Milkyway@home|Sending scheduler request: Requested by user
29/11/2008 21:18:01|Milkyway@home|Requesting 163695 seconds of new work
29/11/2008 21:18:06|Milkyway@home|Scheduler RPC succeeded [server version 603]
29/11/2008 21:18:06|Milkyway@home|Message from server: No work sent
29/11/2008 21:18:06|Milkyway@home|Message from server: (reached per-CPU limit of 5 tasks)
29/11/2008 21:18:06|Milkyway@home|Deferring communication for 1 hr 19 min 21 sec
29/11/2008 21:18:06|Milkyway@home|Reason: no work from project

Looks like I should have looked around instead of jumping right in. The answer to my question is here! Travis lowered the limit about 15 to 20 minutes ago.

I just saw that as well. I wonder if the new work units are long running work units. I guess I won't know until I get one of them on a workstation.

Have the Admins set a new lower WU-per-CPU limit?

I have just gotten this script from my BOINC Manager /Messages/ tab -

29/11/2008 21:17:55|Milkyway@home|Deferring communication for 35 min 8 sec
29/11/2008 21:17:55|Milkyway@home|Reason: no work from project
29/11/2008 21:18:01|Milkyway@home|Sending scheduler request: Requested by user
29/11/2008 21:18:01|Milkyway@home|Requesting 163695 seconds of new work
29/11/2008 21:18:06|Milkyway@home|Scheduler RPC succeeded [server version 603]
29/11/2008 21:18:06|Milkyway@home|Message from server: No work sent
29/11/2008 21:18:06|Milkyway@home|Message from server: (reached per-CPU limit of 5 tasks)
29/11/2008 21:18:06|Milkyway@home|Deferring communication for 1 hr 19 min 21 sec
29/11/2008 21:18:06|Milkyway@home|Reason: no work from project

---

What's up with the new limit? 5 tasks per CPU? You're kidding ,right? I complete 4 units every 5 minutes on my quad. Unless there is a large amount of work units available at all times I don't see how this is going to work. At the 20 per CPU limit I would run out of work very quickly and keep getting the message "No Work Sent"