P-1 Question

[cheesehead]

Originally Posted by NBtarheel_33

Question: What was the minimum typical RAM, call it x, in a computer at the time that the 8 MB threshold was chosen?

"Typical" not the point!!!

It's not the typical (even the "minimum typical") that should govern the default choice, but the (even if rare) minimum systems. GIMPS's goal of minimal interference needs to be geared to the minimum volunteered systems, not the typical volunteered systems.

What is the minimum typical RAM, call it x*, in a computer now?

That's not relevant. No matter how much RAM a system has, GIMPS needs to look at the case where the user already needs most of it for applications other than GIMPS, then minimize its default added burden.

Would it not be fair to claim that we ought to be able to adjust the default P-1 RAM to a value closer to z, where x*/z = x/8?

No, it's not at all fair unless the user has given consent to allocate that much to GIMPS!

... and the user can give that consent by changing the Allocated Memory figure himself from the default value to a higher figure.

Or, if you want to be extra careful, how about taking z to be half of the value that is obtained from the equation x*/z = x/8?

No, that still is not fair without the user's consent as indicated by user adjustment of Allocated Memory from the default 8M to a higher value.

We have to be careful not to get in our donors' way, yes, but

but you want to get more in our donors' ways than we do now.

perhaps there was a time when we risked this more so by blithely asking for 8 MB. Maybe asking for 8 MB back in 1997 is like asking for 128 MB today.

Was it? Please show evidence.

[diamonddave]

Originally Posted by cheesehead

"Typical" not the point!!!

It's not the typical (even the "minimum typical") that should govern the default choice, but the (even if rare) minimum systems. GIMPS's goal of minimal interference needs to be geared to the minimum volunteered systems, not the typical volunteered systems.

That's not relevant. No matter how much RAM a system has, GIMPS needs to look at the case where the user already needs most of it for applications other than GIMPS, then minimize its default added burden.

No, it's not at all fair unless the user has given consent to allocate that much to GIMPS!

... and the user can give that consent by changing the Allocated Memory figure himself from the default value to a higher figure.

No, that still is not fair without the user's consent as indicated by user adjustment of Allocated Memory from the default 8M to a higher value.

but you want to get more in our donors' ways than we do now.

Was it? Please show evidence.

I don't think he meant to change the default on current installation. But to update the default that will be proposed to the user on new installation.

[axn]

Originally Posted by cheesehead

So did I

I'm sorry, but I cannot reconcile that with this

Originally Posted by cheesehead

It isn't helpful for GIMPSters to start exhibiting an attitude that we are somehow entitled to more resources on those systems (that are not ours) on which we have voluntarily been given the use of background time.

If the "we" is referring to the forum's P-1 volunteers, how did you get to general GIMPS participants?

[axn]

Originally Posted by cheesehead

"Typical" not the point!!!

It's not the typical (even the "minimum typical") that should govern the default choice, but the (even if rare) minimum systems.

You'd make a lousy project administrator Risk-benefit calculations and tradeoffs are made all the time. Typical _should_ be the point.

But, I'm not interested in convincing you. George, perhaps.

[Rodrigo]

Originally Posted by axn

You'd make a lousy project administrator Risk-benefit calculations and tradeoffs are made all the time. Typical _should_ be the point.

But, I'm not interested in convincing you. George, perhaps.

I think cheesehead is on to something here.

I may not have much programming or mathematical expertise, but I do know a thing or two about dealing with staffs of volunteers, which may be a better analogy than a project administrator (in a professional setting). The operative phrase in situations where your personnel can leave without penalty at the drop of a hat (since they're not getting paid), is: BEGGARS CAN'T BE CHOOSERS. Make changes and regulations that annoy the volunteers (in the case of GIMPS, that cause noticeable resource drains on volunteered non-dedicated PCs), and I'm going to be left with a considerably emptier office. It's a difficult balancing act.

Sure, I can console myself with the line that the remaining volunteers are "more committed," that their "per-worker output" is bigger, and so on and so forth, but the bottom line is that less ends up getting done.

In professional settings, as the manager you're the customer who's buying the employees' time. But in a volunteer effort such as GIMPS, they're actually the customers, as in effect they are purchasing from you the psychic benefits to be gained from offering their time. Decrease those benefits, and you decrease your sales (volunteer efforts).

Rodrigo

[axn]

Originally Posted by Rodrigo

but I do know a thing or two about dealing with staffs of volunteers, which may be a better analogy than a project administrator (in a professional setting).

I'm not talking about a professional setting. I'm talking about a DC project admin (someone like George). Sorry if I wasn't clear. It was no analogy.

[axn]

Originally Posted by Rodrigo

Make changes and regulations that annoy the volunteers (in the case of GIMPS, that cause noticeable resource drains on volunteered non-dedicated PCs)

You're assuming that which you want to prove. The whole idea is that upping the default value to 300MB is not going to cause "noticeable resource drains". The case has been made why 8MB is overly conservative _in today's environment_. To counter that, you've to demonstrate that a significant fraction (say 3%) of the userbase will be affected. Keep in mind that we can only change the default when users install / upgrade to a yet-to-be-released new version of the software -- which users of older machines aren't going to use anyway.

[cheesehead]

Originally Posted by axn

You'd make a lousy project administrator Risk-benefit calculations and tradeoffs are made all the time. Typical _should_ be the point.

1) Suppose a minimum system has applications that use all but 40 MB of RAM without swapping. Prime95 is introduced, and outside of stage 2 exists peacefully without causing any more swapping. But stage 2 with allocation of any more than the default 8 MB does cause swapping, noticeably slowing other applications. Is that acceptable, in your opinion?

2) Suppose a typical system has applications that use all but 40 MB of RAM without swapping. Prime95 is introduced, and outside of stage 2 exists peacefully without causing any more swapping. But stage 2 with allocation of any more than the default 8 MB does cause swapping, noticeably slowing other applications. Is that acceptable, in your opinion?

3) Suppose a maximum system has applications that use all but 40 MB of RAM without swapping. Prime95 is introduced, and outside of stage 2 exists peacefully without causing any more swapping. But stage 2 with allocation of any more than the default 8 MB does cause swapping, noticeably slowing other applications. Is that acceptable, in your opinion?

[cheesehead]

Originally Posted by axn

I'm sorry, but I cannot reconcile that with this



If the "we" is referring to the forum's P-1 volunteers, how did you get to general GIMPS participants?

If the default is changed for the people who volunteer for P-1, how do we not change the default for general GIMPS participants?

Assume you come up with an answer for that. How does that materially differ from simply having the volunteers each change their allocated memory from the default to 300 MB (or whatever you're proposing)?

In other words, if the default is not going to be changed for general GIMPS participants, how does that materially differ from the situation we already have right now, in which volunteers perform one simple parameter change to their allocated memories?

But if it is going to be changed for general GIMPS participants as well as for "we" volunteers, then -- well, we've just connected "we" volunteers to general GIMPS participants, in regard to default values, haven't "we"?

[axn]

Originally Posted by cheesehead

1) Suppose a minimum system...

All three of scenarios are trivially invalid. Prime95 has a working memory requirement of about 50MB. So merely running P95 will cause trashing. We don't have to get to Stage 2 for that. But I'll assume that there is a certain usage level at which Stage 2 will cause trashing, and it is this that you're interested in (not the specific number 40).

Also, it is my understanding that P95 will not acquire more than 90% of available free memory. So to cause trashing, P95 should've acquired the memory for stage 2 and then additional load from other programs caused the trashing.

Is it acceptable? It is most unfortunate, but if P95 is only using 300MB, I think we have a definite defense -- we are _not_ using too much resources -- for a high end and typical systems.

For the minimal system, depending on how much it is spec'ed, may not find it acceptable. OTOH, it is highly unlikely that these minimal system will have enough specs to get LL/P-1 in the first place.

In short, I am not worried about this scenario happening in real life.

[axn]

Originally Posted by cheesehead

If the default is changed for the people who volunteer for P-1, how do we not change the default for general GIMPS participants?

First off, Dubslow didn't come up with the proposal for change of default -- I did. So perhaps I can clarify certain misconceptions you're having.

People who volunteer for P-1 == those who do so in this forum. That's what I mean -- and that's what I think Dubslow meant when he said "we" -- not the wider P-1 community. If that's not what you mean, we're talking different things. (For those who do volunteer here, the question of default doesn't even arise. They know what they're doing -- they'll allocate necessary memory).

When I say "change the default", I mean, make the _new_ P95 versions such that instead of showing 8MB, show 300MB, when a user installs (or upgrades) P95.

Most importantly, I didn't interpret Dubslow's statement as at all dealing with change of default (his was just a casual support of my idea). His statement was still concerning _his_ original idea of testing partially done (B1=B2) P-1's. That's where I *ahem* accused you of less-than-charitable interpretation.

Originally Posted by Dubslow

I also agree, and it is in that spirit that I originally asked my question, though of course I'm trying to treat the symptoms, not the problem, when we're already short of resources.

(my emphasis)
My interpretation -- he kind of admits that doing the partial P-1 is a lost cause, since there is not enough resources to even do a proper P-1 on "virgin" candidates. Seriously? Is that not a viable interpretation?