Hmm, I\'m currently crunching on a workunit that I just got yesterday, still have over 1,100 hours to go, a little over 2% complete but I\'ve now got 160.68 credit and it\'s for the workunit I\'m doing now. Can anyone tell me why that happened if they know as that seems very unusual, thanks.

Hmm, I\'m currently crunching on a workunit that I just got yesterday, still have over 1,100 hours to go, a little over 2% complete but I\'ve now got 160.68 credit and it\'s for the workunit I\'m doing now. Can anyone tell me why that happened if they know as that seems very unusual, thanks.

Nevermind, I apologize all, I missed the part in the FAQ about how it gives a report every 10,802 timesteps and you\'re given some credit at that point.

Welcome to cpdn.
It\'s a long, slow journey to the finish, but if you have problems, just post on the help boards.
There\'s always someone around to give advice.

Yes, CPDN is different. But, there are other projects in the pipeline that have suggested having work units that take months to run also, so, lets be glad they pioneered this change ...Otherwise I would donate 90 days time before I would see credit?

Not likely ... :)

I want to do the science, which is why I am here, but I like to get \"paid\" for the effort too ...

---

Yes, CPDN is different. But, there are other projects in the pipeline that have suggested having work units that take months to run also, so, lets

HA, yeah, when I first started there was no means for credits/stats etc until the end of a run -- which was a 3-6 month job at the time, so I came up with the \"trickle\" for credits & basic stats, and now it\'s doing much more. Oof course with \"Moore\'s & Murphy\'s Law\" we\'re getting back to 3-6 months workunits with spinups & coupled models! ;-)

PS -- Do you know which projects are going to have months-long workunits a la CPDN, I\'d like to email them my condolences now! :-)

Folding@Home and The Lattice Project are the two that come to my mind.

One of the alpha testers can probably confirm this better than I, especially as I don\'t run any test projects (I have barely enough time to handle the problems that come up with just the normal projects).

In any case, I have not heard anything that suggests that they are anywhere close to moving to production.

---

FaH will have longer than normal workunits but still not on the scale of CPDN. Also the BOINC version of FaH will encourge \'timeless\' workunits, these are designed for slower computers so they tend to be short for FaH with longer deadlines.

I hope Lattice does not have workunits this long, as far as I know they still have not figured out how to checkpoint so it could get really ugly if they do.

---

BOINC WIKI

BOINCing since 2002/12/8

Yes, CPDN is different. But, there are other projects in the pipeline that have suggested having work units that take months to run also, so, lets

HA, yeah, when I first started there was no means for credits/stats etc until the end of a run -- which was a 3-6 month job at the time, so I came up with the \"trickle\" for credits & basic stats, and now it\'s doing much more. Oof course with \"Moore\'s & Murphy\'s Law\" we\'re getting back to 3-6 months workunits with spinups & coupled models! ;-)

PS -- Do you know which projects are going to have months-long workunits a la CPDN, I\'d like to email them my condolences now! :-)

Well I have no problems with just getting the credit at the end, wasn\'t expecting anything different here, mostly because I didn\'t thoroughly read the FAQ. Almost had to post about the Sulphur workunits as well as I read the time estimates before I found out about the Sulphur variants and wondered why the estimated time for completion was about double that of a slower computer in the FAQ haha.

As far as which ones are going to long workunits I\'ve heard that there is a Alpha testing SETI@Home going on right now. I can\'t recall the length of the workunits that was mentioned though so I can\'t say how it compares to CPDN.

The SETI@Home enhanced work units will be from 4 to 10 times longer than present. For me that means 12 to 40 hours with an non-optimized science application. The phase in will be gradual as they only release some work so that the load on the download server will be managable with each new person needing to download the new science application.

Einstein@Home on the the other hand has gone from 11-12 hours to 7 or so on Windows and as low as one hour on my PowerMac G5 ...

In part this is because they do have an Altivec enabled science application for OS-X but not yet one for Windows, or, because of the differnces between AMD and Intel chips have not been able to come up with one that is satisfactory (returns valid results), doesn\'t crash, and is significantly faster than the non-optimized application.

So, time wise, SETI@Home and Einstein@Home are exchanging places ...

---

Well I have no problems with just getting the credit at the end...

I would. Suppose the WU dies. Classic was good at returning info from a broken WU and kept a separate count of \'short\' wu. But suppose for some reason the wu cannot be returned.

On the one hand the science has no data, you can say then that no credit at all is fair.

On the other hand the user has made the effort, you can say full credit is fair.

But in between those two extremes is something that is even fairer than either - a compromise that relfects both ways of thinking. Interim credit for interim check-ins, which you don\'t lose even if the model fails - and a bonus for finishing as that is what is useful for the science (if I remember right this is paid as a larger than normal credit for the last trickle).

And that is what CPDN came up with. In my opinion it is a brilliant extension to the BOINC framework. Even if the projects Paul mentioned do not use it, someone ele someday will need it.

River~~

---

The big bonus at the end was what caused a large amount of grief, as only some people got it.
It took Carl a long time, a lot of work, (and more grief for some), to come up with a method whereby NO ONE got a \'large bonus at the end\', just the usual amount for the final trickle.

LHC@Home has tried to do a partial reward. BUt, I don\'t think it works as well as it should. In part this is because their validation requirements are so strict that one digit off at the 16-18th decimal place is tough ... :)

You do have a point, I put in x computional effort and yet failed through no fault of my own it is discouraging to say the least. I had a burp with Rosetta the other day and lost a days worth of work because of a slip-up with preferences. So, should i be rewarded or not? :)

If only there were not those that would take advantage of the honor system. Ah,well, assuming the system are equivelently stable we should see the same distributions of work unit failures over time. Yet there are those models that die because of a flaw in the simulation or the code. I have had a few CPDN models do this. And there are things like I have never been able to run a model successfully on the G5, they all crash sooner or later ... :(

I could almost see two accounting systems where you have effort expended, vs. effort expended that results in science. In theory the two should be identical or close to identical ... but, I would rather we get the basic credit system back to some semblance of fairness ...

There was a question raised about a G5 system that benches at 11K IntOPs, yet, its processing time is the same as my G5 which does 1100 IntOPs ... unknowing use of a \"superbench\" client, outright cheating, or something else? This harkens back to the discussion where the asertion is made that optimizing the benchmark only makes it report what it \"should\"... :)

oh well ...

Yes, the simplicity of the CPDN system has something to recommend it ...

---

Well I have no problems with just getting the credit at the end...

I would. Suppose the WU dies. Classic was good at returning info from a broken WU and kept a separate count of \'short\' wu. But suppose for some reason the wu cannot be returned.

On the one hand the science has no data, you can say then that no credit at all is fair.

On the other hand the user has made the effort, you can say full credit is fair.

But in between those two extremes is something that is even fairer than either - a compromise that relfects both ways of thinking. Interim credit for interim check-ins, which you don\'t lose even if the model fails - and a bonus for finishing as that is what is useful for the science (if I remember right this is paid as a larger than normal credit for the last trickle).

And that is what CPDN came up with. In my opinion it is a brilliant extension to the BOINC framework. Even if the projects Paul mentioned do not use it, someone ele someday will need it.

River~~

Well the credit being all given at the end of a workunit doesn\'t bother me because I\'m doing it for the science, not the credit/ranking involved. Each person does it for their own reasons which doesn\'t matter so long as the science gets down, so long as we all arrive at the same destination our reason for going there doesn\'t matter.

However, I do agree with you, the way CPDN gives credit really is a outstanding way to do it, hopefully more BOINC projects adopt something similar for their projects.