MW work fetch policy

[Rocoll]

I think it's our fault. Not enough WUs for the demand.
When the store of WUs on the server is low it sends a message to restrict WU numbers and frequently sends none which progressively increases the time between updates.
Ultimately you end up with the maximum cycle of just over 3 hours without new units.

Did anyone tell Travis we were doing an AA?

[Furlozza]

There are, of course, two ways to look at the fun and games.
1) Everyone is in the same boat
2) Some folks will drop out, thereby freeing up some additional units..... in all likelihood for cpu units, not GPU
3) There's always WCG biggrin

[yoda]

Did anyone tell Travis we were doing an AA?

Not me.  But then, in the overall scheme of things, the impact we are having is relatively small anyway.  If you sort by the credit for the week (at BOINCStats), the top 10 teams earned 85.8 million credits between them.  Only 3.8 million of that (less than 5%) was done by our team.

[clownius]

Im guessing MW is taking an ATI hammering lol.  Thats what happens when u have the only project that can use a type of GPU

[Mike Mitchell]

What are people using as their connect to and work buffer policies?  I just changed mine from connect every .25 days and a buffer of .25 days to connect every 0 days and a buffer of .2 days to see if it helps.

I rest mine to 0.0 and 0.25 and that didn't make much difference, if any. I've added a few backup projects which seems to have helped. MW runs dry for a while, then downloads a large batch. So it suits me. The backup projects I used were my lowest ranked ones. MW will make up for the lower credit.