Perfect timing.

You may be missing the point which is that there are about 89,000 active users today. If each of them do a manual update several times each day, that makes the servers work much more than if you keep you hands to yourself. You do the math: 89,000 times just .1 second each = 2.472 hours just to acknowledge the last update request.

Plus you are wrong about the 5 day cache. The client will auto update when the cache falls below 5 days. If you are currently crunching 6 hour WU, then the auto update will normally run about 4 times per day, all without you intervening and stressing the servers.

Hence Sekerob's fervent request: Step away from the Update button.

nasher, work requests are almost always before the RtR... backfill of cache / supplying new work has higher priority than clearing RtR, so sequentially:

- Cache-up
- Nearing point where job is finishing so new cache-up
- Work completes and uploads
- 1 RtR sits there
- Nearing point where job is finishing so new cache-up, and oh, yes good time to report the RtR.
- Work completes and result files upload
- 1 new RtR sits there
- Suddenly, extra long job, finishes and cache inflates and is now worth 6 days
- Work completes and result files upload
- 1 RtR still sits there and not 24 hours old and new one is added, now 2 sit there.
etc etc etc

Communication is not merely limited to setting the cache. You also need to set the Connect Every X days and of course set the Activity Menu to "Network according preferences", else when a intertube line is there, the client will just do what it does and assumes it's always allowed to work through the report/fetch sequence, making sure enough work is there to meet His Crunchship's wishes.

WCG default is 0.1 connect and 0.3 days buffer last I looked. Reviewing the statistics, I think this extended outage took out quite a bit of production, so common sense is that Murphy may always visit again and WCG exclusive crunchers would thus do well to pump the cache up a bit. My semi regular pref is 0.9 days. Never run on empty and the new client 6.10.x backup project function only going out to alternate projects with a zero weight setting until the cache is really bone dry. I've got 3 because neither of these have a constant supply of work.

ON topic, 88:16 hours per midday stats and 76 hours in PV... Ruby Tuesday is still some months away... first some blue approach goals to fill to at least bring all the researches to a point if when a 30 day end is announced the ability is there to flip the switch and get to that target well within that time...within 15 calendar days is the target. Limited resources require turning on 'managed' crunching

Based on the flows recorded for past 6 days I'd say we'll be seeing non stop supply at this pace for at least the next 10-14 days.

12th 4306
13th 13723
14th 18013
15th 20229
16th 26851
17th 28434 (projected)

That's 111,000 validated (for 36 CPU years), and another half in PV (big thumb). How much in client caches is anybody's guess, for the techs to know. Uplinger wrote in that he's releasing additionally 2x83,000 which then also comes with about 4-5% in make-ups. In all should things dry up with this ts run, we'll probably have done 125-150 CPU years worth, very ballpark. As for 'ts', don't know if the full sweet is run to simulate a complete target. Very slowly we're creeping towards the million results as was forecasted for a single target.

Sit back, relax, DDDT2 is not going away.

I do not know that I agree with you that there is a "non-stop" supply!

I am getting the following message frequently on my 20 Windowss machines:


8/17/2010 4:24:10 PM World Community Grid Message from server: No work can be sent for the applications you have selected
8/17/2010 4:24:10 PM World Community Grid Message from server: No work is available for Discovering Dengue Drugs - Together - Phase 2 (Type A)
8/17/2010 4:24:10 PM World Community Grid Message from server: You have selected to receive work from other applications if no work is available for the applications you selected
8/17/2010 4:24:10 PM World Community Grid Message from server: Sending work from other applications



It looks like only 50% of the time I am getting DDDT2 WUs even though I only have DDDT2 as the only project selected.

On the other hand, for my 5 Ubuntu machines with CEP2 selected as the only project, 100% of the WUs are CEP2.