Paul, there are a lot of valid issues in your message, and I agree with you on them. You're in a tough situation where your connection times are limited and you're suffering under the constraints of the current client scheduler and project settings.
First off, the devs seem to be aware of the WU cache vs. connection time difference. Most of them, anyway. :) The two are not always the same, and for people on dial-up or intermittent network connections, there are important distinctions to be made between them. Hopefully we'll see some changes for the better on this topic come about.
Second, the accurate estimation of WU processing times has become vital to the successful scheduling for all project work. Bad estimates on how long it takes to actually crunch lead to bad scheduling decisions on what to work on next, and how much work to grab when it goes looking for more. This is hugely complicated since not all the projects have WU's of similar size or that contain an equal amount of work. The devs are brainstorming solutions for weighted averages on actual completion times per project, and other options that might produce some big changes for the better here.
My fundamental point was that, although 4.45 can require a shift in assumptions on what a good connect every / cache size value should be, it works. It might not be perfect, but it works. Some people might not agree with what they see happening at any given moment, but work is getting done. (Actually, the new core clients do give you better control over your projects and work, so micro-managers should rejoice!) For those who don't like the way that it works, there's 4.19 to fall back on. It doesn't sound like they're going to take that away from us anytime soon. ;)
Paul, have you tried using the new Mac application that Bruce put up yesterday?
No,
My luck with doing those things is that all I wind up doing is corrupting my BOINC and client erroring all the work. I have tried it with SETI@Home a couple times ... and with the work here taking so much longer ... well ..
Besides, I have enough trouble just doing my other stuff ...
Paul, there are a lot of
)
Paul, there are a lot of valid issues in your message, and I agree with you on them. You're in a tough situation where your connection times are limited and you're suffering under the constraints of the current client scheduler and project settings.
First off, the devs seem to be aware of the WU cache vs. connection time difference. Most of them, anyway. :) The two are not always the same, and for people on dial-up or intermittent network connections, there are important distinctions to be made between them. Hopefully we'll see some changes for the better on this topic come about.
Second, the accurate estimation of WU processing times has become vital to the successful scheduling for all project work. Bad estimates on how long it takes to actually crunch lead to bad scheduling decisions on what to work on next, and how much work to grab when it goes looking for more. This is hugely complicated since not all the projects have WU's of similar size or that contain an equal amount of work. The devs are brainstorming solutions for weighted averages on actual completion times per project, and other options that might produce some big changes for the better here.
My fundamental point was that, although 4.45 can require a shift in assumptions on what a good connect every / cache size value should be, it works. It might not be perfect, but it works. Some people might not agree with what they see happening at any given moment, but work is getting done. (Actually, the new core clients do give you better control over your projects and work, so micro-managers should rejoice!) For those who don't like the way that it works, there's 4.19 to fall back on. It doesn't sound like they're going to take that away from us anytime soon. ;)
Now, if only they would
)
Now, if only they would finish and compile the 4.45 or 4.46 for Mac ...
Not that I am seeing that bad of a problem with 4.43 ...
Paul, have you tried using
)
Paul, have you tried using the new Mac application that Bruce put up yesterday?
RE: Paul, have you tried
)
No,
My luck with doing those things is that all I wind up doing is corrupting my BOINC and client erroring all the work. I have tried it with SETI@Home a couple times ... and with the work here taking so much longer ... well ..
Besides, I have enough trouble just doing my other stuff ...