Late hand-ins still useful?

[stoffer]

Dear Folding Home,

I am very impressed with the project, and I would like to help with what I can. However, I have some trouble getting workunits done for the deadline on my relatively new C2D iMac unless I let it run 24/7 with no power-save options.

While I can understand the point reward system drives many users, I care little about it - so all I want to know is wether workunits handed in late are still useful even though they reward no points?

If late hand-ins are useful, I will keep folding - if not, its waste of energy.

Please let me know (:

Thanks!

[MtM]

which deadline: preferred or final? Which client?

Preferred deadline, if you miss it you get less points but still some..
Final deadline, if you miss it you won't get points, but more importantly the server will issue the same wu to another donor so you'll be doing duplicate work which is not usefull.

Btw, I'm pretty sure you will be able to complete units in time, maybe not smp units but for the classic client the deadlines are leniant enough that people with older computers can turn in work in time. So a c2d should have no problems with it at all, unless it's maybe only on one hour a day.

Edit: btw, the SMP client was intended for quad core machines, not dual cores. That people were able to complete units in time on them was because of the unit's being small enough and the machines being often dedicated folding machines. If you don't run it as dedicated folding machine, I would not run SMP on anything less then a fast quad as the deadlines are key to the project.

[lanbrown]

Once the preferred deadline is past, the WU is reissued to someone else. You will still get credit for it as long as it is returned before the final deadline. If you are consistently not meeting the preferred deadline, then you really should look at the client settings. Chances are, you are doing SMP and you would be better off running two clients that utilize a single core. The deadlines are much further out and you won't have an issue not meeting the preferred deadlines. Under the current client, you should stop it and add -oneunit to the client and let it run. Once it is done. it will upload the WU and not get another. Then you can reconfigure the client with the -configonly flag and then add -smp 1 to the parameters so that it will only use one core. This will get you non-SMP WU's, see how it runs for one WU and if all is well, you could add a second client with the same settings (the machine ID will need to be different though) and now you are using both cores.

[MtM]

Does -smp 1 give you classic wu's, or would you need to remove -smp x completely

Also, there is a reason there is still a classical client on the download pages, and not only the smp client which addmitedly can act as a classic client when you ommit the -smp flag. This reason is that the classic client is fully developed, set it and forget it. The smp client isn't yet as stable, even when running in 'classic mode' you might be vulnurable to the odd bug here and there.

When OP posted back that he has an SMP client, I would have adviced him to stop it, add -oneunit, check if the unit is send after it's finished and then set up the normal classical client times two.

I did goof up with saying the unit is reissued when you pass the final deadline instead of the preferred though, thanks for the correction.

[lanbrown]

MTM,

He does have a Mac and I'm assuming he is running the OS X client which has been SMP for awhile. The classic client is no good for him (PPC client that would run under Rosetta) and I believe even it was SMP enabled. He needs to use the current client and -smp 1 will make sure that only one core is used. The assignment server should take that into account and assign him to a classic work server. The OP could try without the smp flag and see what it does.

[patonb]

Nope, as soon as you add smp, it gives a smp unit and not a unicore. The smp x settings have no bearing on what wu you get.

You just don't add the smp flag at all for win machines......... atleast.

[MtM]

Ian sorry, you're probably right, I must have overlooked the 'iMac' part. I'm not sure how the Mac client works now, I know it's smp capable, and I know you can run the windows smp client as classic client by ommiting -smp x totally. Maybe for mac the behaviour is diffrent I don't know.

[stoffer]

ok - I just got myself the newest one that says SMP as you mention. I fold 5x faster than deadline if I leave it on 24/7, even with use during the day. I prefer to let it sleep at night though etc etc.

So, I turned FAH off for now, so no energy is wasted. Does anyone have solution to let the mac client request smaller WUs?

Thank you!

[MtM]

Allow the reciept of work units ( small/normal/large ) : small

Do not say yes to advmethods.

I can't make sence of what you're saying though, I run a quad and have +- 90% of deadline remaining running 24/7, you're saying you got 80% of deadline remaining with your c2d iMac. Could be, if you have a very fast wolfdale ( I have an old kentsfield clocked stock ).

Could you post some of the log file in here, preferable between

Code: Select all

[code] 

[/code] tags?

[stoffer]

Working in windows boot camp right now - when I boot back again, I will post from the log.

well, I have a 24 inch alu iMac with C2D from 09!? and if I leave it be with power saving mode off, it reports WU's done with about 5 times deadline speed.

I will report back!

[PantherX]

Last time I checked, Macs are only capable of folding SMP WUs. There isn't a Classic Client or WUs for Mac.

@lanbrown: Can you please tell me where you came across -smp 1 because I was informed by a well known user that -smp 1 isn't a valid argument (and that is correct). From my personal experimenting, I found that -smp 1 = -smp so if you are interested in the details, here is my post.

[stoffer]

then I (and FAH) is out of luck with work done from my mac I guess ):

[PantherX]

Well, I suggest that you use the SMP Client but with -oneunit flag so whenever it finishes, your system can power-off (please note that I am not a Mac User) and when you feel like folding again, start the Client and it will download another WU and repeat the cycle. That way, your system won't fold 24/7 but only when you start the Client, it finishes and exits.

[MtM]

I don't know about that, if he has a c2d it will most likley take longer to do an smp work unit then the time he will have his computer on. So the unit might not finish in time, in which case he would be back to square one?

Edit for bad sentence.

[7im]

Allow the reciept of work units ( small/normal/large ) : small

Do not say yes to advmethods.

I can't make sence of what you're saying though, I run a quad and have +- 90% of deadline remaining running 24/7, you're saying you got 80% of deadline remaining with your c2d iMac. Could be, if you have a very fast wolfdale ( I have an old kentsfield clocked stock ).

Could you post some of the log file in here, preferable between

Code: Select all

[code] 

[/code] tags?

Currently, the only client available to the Mac is the SMP client. On a 2 core system, it will need to run most of the time to complete the WU before the deadline.

Selecting a smaller WU size will not help. SMP is hard coded to Large WUs. Not selecting -advmethods won't change anything related to WU size either.

PantherX had the best suggestion. Start a WU in the morning when you turn on the computer, and then let it finish the WU with the -oneunit option. Then set you mac to sleep when it becomes idle (after the Fah client exits). Hopefully, the WU will finish before morning, and won't run all night like you wanted.