I was going to have a big long speech about this but I will shorten it up.

I do not like the "time slicing" feature. I want to participate in multiple projects and only get so many wu's that I designate them to get from one project at a time. I don't want a mixture of wu's from the different projects.

This might be a good feature for some in ones own global preferences, but it should not be automatic for everyone.

Anyone else want to rant on this?

Just curious.

CPDN work units take several weeks, vs. 2 or 3 hours for S@H.

So, if you want to divide equally and not time slice, you'd do one CPDN work unit, then spend 3 weeks exclusively on S@H, then one week on CPDN....

I think that's the logic behind slicing -- to deal with the disparity between projects like SETI and CPDN....

> I do not like the "time slicing" feature. I want to participate in multiple
> projects and only get so many wu's that I designate them to get from one
> project at a time. I don't want a mixture of wu's from the different
> projects.
>

But that's not how it works now. Right now I have a CPDN and a Seti wu and they split time, which I don't want to happen.

You've signed-up for two projects - presumably because you wanted to run both of them....

How else would you do it?

At the moment, depending on the ratios you've set for the projects, you're probabbly running several SETI units, then switching to CPDN for an hour or so, then back to SETI

What's the problem?

Perhaps the big long speach might clarify your concern...


> But that's not how it works now. Right now I have a CPDN and a Seti wu and
> they split time, which I don't want to happen.
>


>
>

You can sign up for several projects, but each host (pc) will only run the projects its "attached" to - so attach one to seti, one to cpdn, one to predictor - and others to combinations. Then only the one(s) with more than one attached project will timeslice.

M.

They way it use to run was like this: I would get 3 or 4 Seti wu's and crunch them all before I got any Predictor wu's.

--- - 2004-09-01 09:08:38 - Insufficient work; requesting more


I would get a message above when I only had a few more hours left on a Seti wu before it downloaded Predictor wu's. I have weeks to go on my CPDN wu so I don't need any Seti wu's at the moment.

I want to finish my CPDN wu, then do Seti, then Predictor. I don't want one of each because the CPDN takes so long.

> I want to finish my CPDN wu, then do Seti, then Predictor. I don't want one of
> each because the CPDN takes so long.

CPDN does take a long time, but you get credit every few hours of processing it does - called 'trickle' credit.

So, you should consider one CPDN unit as several hundred smaller units.

I know about the trickle credit but it isn't several hundred units.

If you don't know why I don't like this, then I can't change your mind about it. I like BOINC and I like the projects but I don't like this feature. And I know all about the resource share crap too, which doesn't work yet. It still changes every hour.

I guess I have three choices: 1. accept this change, 2. detach from the project or 3. delete the seti wu from the project folder.

Ah ok agreed - It is not perfect. But it does work - for me at least.

I don't see hourly switching, but then my ratios for SETI:CPDN:Predictor are 75:20:1. As such, I'm running 3 or 4 SETI units then a slice of CPDN (Predictor is n/a for V4 at the mo).

Personally, I would like to see the concept of 'long' and 'short' projects, with short units running to completion, never preempted. Long units should be run either to their 'trickle' point, or if too lengthy, to their disk checkpoint.

Here is what mine looks like: (resource share of CPDN 200 and Seti 100)


2004-09-01 11:57:33 [climateprediction.net] Preempting result 010c_100026293_0 (quit)
2004-09-01 11:57:34 [SETI@home] Restarting result 04my04aa.25870.22257.679814.101_1 using setiathome version 4.03
2004-09-01 12:57:35 [climateprediction.net] Restarting result 010c_100026293_0 using hadsm3 version 4.02
2004-09-01 12:57:35 [SETI@home] Preempting result 04my04aa.25870.22257.679814.101_1 (quit)
2004-09-01 13:57:35 [climateprediction.net] Preempting result 010c_100026293_0 (quit)
2004-09-01 13:57:35 [SETI@home] Restarting result 04my04aa.25870.22257.679814.101_1 using setiathome version 4.03
2004-09-01 14:57:36 [climateprediction.net] Restarting result 010c_100026293_0 using hadsm3 version 4.02
2004-09-01 14:57:36 [SETI@home] Preempting result 04my04aa.25870.22257.679814.101_1 (quit)
2004-09-01 15:57:36 [climateprediction.net] Preempting result 010c_100026293_0 (quit)
2004-09-01 15:57:37 [SETI@home] Restarting result 04my04aa.25870.22257.679814.101_1 using setiathome version 4.03

I don't know if this would work, but how about disabling BOINC's network access and doing manual updates peridically. That way you can finish crunching the Seti WU's and then just let it crunch away on CPDN WU's.

> I was going to have a big long speech about this but I will shorten it up.
>
> I do not like the "time slicing" feature. I want to participate in multiple
> projects and only get so many wu's that I designate them to get from one
> project at a time. I don't want a mixture of wu's from the different
> projects.
>
> This might be a good feature for some in ones own global preferences, but it
> should not be automatic for everyone.
>
> Anyone else want to rant on this?
>


>
>

Easy solution, signup for ONE project problem solved.....



M7 Seti@h Berkeley's Staff Friends Club ©

I like the time slicing, if it wasn't available I wouldn't run CPDN - I don't want to dedicate my machine to one project for that long a period of time. I want to do some of everything.

Personally, if I was to ask for a change it would be that I had the ability to assign each of my virtual processors on my HT machine it's own task so that I would be able to dedicate one processor to CPDN and let the other processor crunch SETI and Predictor units. But I'm still happy with the way timeslicing is working out.

What the hell's a rant?
____________________________________________________
"To alcohol: The cause of, and solution to, all of life's problems!" - Homer Jay Simpson

> But that's not how it works now. Right now I have a CPDN and a Seti wu and
> they split time, which I don't want to happen.
>


>

What's the difference?

33% of the time will be 33% of the time - one way or another - over time.

Similar to: If you want your child to walk as soon as possible, you can coach and cajole and hold them up and they will almost certainly walk by the time they are 12 months old.

If you just let them walk when they are ready, they will almost certainly walk by the time they are a year old.

> What the hell's a rant?
> ____________________________________________________
> "To alcohol: The cause of, and solution to, all of life's problems!" - Homer
> Jay Simpson
> <img> src="http://boinc.mundayweb.com/seti2/stats.php?userID=1175&trans=off">
>
rant Audio pronunciation of "rant" ( P ) Pronunciation Key (rnt)
v. rant·ed, rant·ing, rants
v. intr.

To speak or write in a angry or violent manner; rave.


v. tr.

To utter or express with violence or extravagance: a dictator who ranted his vitriol onto a captive audience.


n.

1. Violent or extravagant speech or writing.
2. A speech or piece of writing that incites anger or violence: “The vast majority [of teenagers logged onto the Internet] did not encounter recipes for pipe bombs or deranged rants about white supremacy” (Daniel Okrent).
3. Chiefly British. Wild or uproarious merriment.

> > What the hell's a rant?
> > ____________________________________________________
> > "To alcohol: The cause of, and solution to, all of life's problems!" -
> Homer
> > Jay Simpson
> >

I don't like the time slicing.
I had 5 LHC wu all with the deadline 2 days past.
One of these wu was in the preempted stage with one minute left to complete.
I would prefer to have projects with the following options.
Project A is the number one project as long as there is work that is the one that get completed.
Project B if project A runs out of work than project gots done.
Also watching the deadline on each project.
project B,C,D etc would each have a wu on the computer.
when deadline approches the wu being work on would be completed before switching.
The object is to work on the project I want to work on with a backup projects.
I had sign up for "find a drug project."
It has many major app. Cancer, Respiratory Diseases, Proteome, HIV, Multiple, Sclerosis, Bioterrorism Antidotes and Methodology.
You could not choose witch on to work on, it assign one to you.
If I am going to donate my resources I want it to work on the project or projects with the piority that I want it to use.
I deleted the "find a drug project" from my computer.
I hope this is long enought.

Fred BOINC Alpha & BOINC Beta Tester

> It was a joke. From Family Guy. Let me see if I can find the exact quote.
> Ahhh...here it is:
> "Dennis Miller(on TV): I don't wanna go on a rant here, but American foreign
> policy makes as much sense as Beowulf having sex with Robert Fulton at the 1st
> battle of Antietam. I mean when a neo-conservative defenestrates, its like
> Ramonsikov filibuster deoxymonohydroxinate.
> Peter Griffin(watching):What the hell's a rant?"
> Ahhh...nevermind

I'm in New Zealand...so "cultural" references to Television programs I've never seen will almost always be unrecognised and incomprehensible.

To make matters worse, I don't watch TV anymore. I've seen all the plots, laughed at all the jokes...and for news TV is worse than a waste of time. They tell lies on TV. :-)

( Not quite a rant.....but close! )

It would be nice if they had more control users could exert on BOINC.



Project preference percentages do seem really weird....if you set it for 99% for 1 project and 1% for another it seems to do each project 50% of the time. which makes no sense at all...you would expect 1% of cpu time on one project and 99% on the other.


It defeats the whole purpose of having project sharing if BOINC won't let the users implement it in a rational way...saying to a person to just run 1 project on each computer to get around BOINCS peculiar project sharing is redundant to having project sharing on BOINC in the first place.


A more rational way of using the resource shares we set and allowing users to input exceptions and special cirmcumstances BOINC should operate under would go a long way making alot of us happier.