• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

To fold or not to fold?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

monkeyclaw

New Member
Joined
Feb 27, 2014
So, I have been wondering about this for a while.

But before I ask, let me briefly introduce myself. My F@H username was started by my brother many years back, probably close to the start of the F@H project itself. Since then it has only been me, my brother, and occasionally one or two of his friends helping us fold. Despite this, our machines through the years have been folding 24/7 whenever we weren't using them, and that has kept us doing reasonably well in the leaderboards and amount of scientific research completed for Stanford :)

So, onto the question:

Sometimes when I know I am about to play a game, and I see that a GPU work unit is about to finish, I just set the GPU slot to finish and let it complete before I play. This way, Stanford gets back its WU faster and I get the best QRB bonus, and everyone is happy.

BUT, sometime there are days when a unit is about to finish, but I probably won't be playing anything for 20 or 40 minutes. And that gets me thinking... should I just let the work unit finish anyway, and hold off downloading that next unit until I'm done with a gaming session, one that could last anywhere from an hour to several hours?

This leads to two inquiries, the first being: would it be considered a folding sin to not fold on a GPU WU for a period of time to avoid downloading a unit and starting the QRB timer? This is the first and primary question.

Secondly, if it is not considered hurtful, and rather would be better for Stanford as well if they receive a WU as quickly as possible after having been assigned...: How complex would be a formula that can weigh out when to download a work unit you know you have to pause in X amount of minutes for an X period of time, only to resume later, VS waiting to start a new unit and not having that head-start, but also not getting the QRB hit of having to pause the WU?

This has been a sort of long and strangely-worded post, but it's just something I've been thinking on, and a question I'd like to bring to the folding community I've been folding for all these years :) Thanks!
 
If you're concerned about some personal commitment issues, there's no answer anyone can give you.

When I want to use Unit #1 to game or other, I simply stop the client and do what I want to do. Sometimes it can be for hours, usually not. But I'm o.k. with it. I'm volunteering my system and Stanford is, in my mind, getting a pretty fair shake out of the deal. Other Folders have dedicated rigs for Folding, or multiples, but they make their own decisions as well.

WTF !!!!!!
 
If it was me, I would just pause the slot and game.

TBH I think Stanford would be happy with any WU being done.
To me it's not about the point; it's all about the WU's. As long as I get the WU in on time, I'm happy. :)
 
Folding is suppose to be donation of idle time so that means first you do what you personally want or need to do, then you let your machine go idle and the folding begins.

Of course its nice if you can wait a few minutes for a WU to finish, but if its inconvenient, I don't think its a folding sin to be selfish.

The only real folding sin is dumping a WU before it finishes in order to get another WU of higher point value.
 
Yeah I agree with what's been said so far. I just think it would be really neat if there was a way to calculate when to fold and when to wait to fold to get the best PPD... Someone with a more robust background in math could maybe tell us what kind of formula/values would have to be known to accomplish this.

I am also doing this mostly because of the benefits that come from the research, but it is really nice to see those points go up after years of paying $30+ a month in electricity for the computer alone :D
 
Basically what you do is fine, just no need to stop it from downloading the next WU. When you want to game just pause it. As long as Stanford gets the WU before the deadline they are happy. The points are just an incentive to keep people folding.
 
Well, my intention is not so much to stop it from downloading the WU, it is to maximize PPD. Points are indeed just points, but incentives are nice and people would never talk about PPD if it wasn't a fun thing to pay attention to, so this post is just exploring that element, all of my WUs are back well before the deadline, being that the machine is a 24/7 rig.

I probably already made my question clear, but I'll illustrate exactly what I mean by an example:

Situation 1: I don't set a WU to finish, and after it finishes the new WU starts and runs for 10 minutes, which I then pause in order to game for 4 hours. I finish gaming and resume folding. This WU now gets a much lower QRB bonus because it was downloaded 4 hours and 10 minutes ago, and only 10 minutes of work is already done on it. PPD is not maximized.

Situation 2: I remember to set it to finish. I start gaming 10 minutes after the WU finishes. I turn GPU folding back on at the end of my gaming session. New WU, fresh QRB timer. PPD is maximized.

Situation 3:
A WU is just about to finish, but I don't quite feel like gaming yet. Maybe in 45 minutes I will, and let's guess the WU that I happen to download next takes a total of 6 hours to finish, and my gaming session will last 3 hours. Is the PPD maximized by starting the WU now and getting that 45 minutes of work in followed by 3 hours of idle time, or should I let the WU finish and then start the WU fresh when I am through?

Situation 1 and 2 would be pretty clear cut, but what about situation 3? Is the rate of the QRB degradation low enough for a WU that size that you should go ahead and start? Where exactly is the cutoff? How hard would it be to make a spreadsheet/formula that calculates this out for each WU?

This was mainly a thought experiment, and I am probably wasting everyone's time here. These are just the things I think about in as a 2-machine-folding-user that is among users that have multiple machines and many, many more people folding on a single username :)
 
Last edited:
Well if you want maximum ppd, then you want the WU to start and finish with no interruptions. Consequently you want to wait until a WU is finished and prevent any new WUs from being download until you can again dedicate your computer to folding.

You would have to calculate the optimum times for each WU. But the penalty for each hour of delayed return is substantial so you don't have a lot of time to game before the penalty grows to unacceptable limits.
 
Depends on the deadline, if you have 1 day from download time to finish it, then let it finish. But if you have 30+ days to turn in a WU, then pause it, game for as long as you want then unpause it :)

Look at the download times and the deadline and you'll see what I mean.
HFMWU.jpg

P.S. I hate core 17 WU's, takes 3 days for my GTS 450's to finish a WU :mad:
 
For shorter gaming sessions, 45 mins or so, I generally pause those.

But like you for uncapped multi hour ones I try to time the Finish to game start time.

I would have thought the needed figures are fairly handy and would not be that hard to make something like you describe.
- Does HFM list what we need?


Sorry, a bit short on time to dive into this at the moment.
 
For shorter gaming sessions, 45 mins or so, I generally pause those.

But like you for uncapped multi hour ones I try to time the Finish to game start time.

I would have thought the needed figures are fairly handy and would not be that hard to make something like you describe.
- Does HFM list what we need?


Sorry, a bit short on time to dive into this at the moment.

HFM does give the deadline, so a short deadline you may want to have it finish the WU. But for long deadlines, just pause it then unpause when your done gaming.
 
Back