Dear all,

Yes, we definitely have noticed the huge increases in credit given out over the weekend. We plan to fix the validator/credit granting system soon and correct recently given credit numbers (sorry, we can't really let you keep those extra credits; the graphs linked above should be more than enough explanation)!

As to why, Thyme Lawn seems to have spotted the reason. We're aiming to migrate away from the assumption that all work-units are the same length (they're not, and this has frequently led to a factor 2 or so misestimate of expected work-unit run-time). Unfortunately though, it appears our first attempt at this has gone wrong. Work-unit run-time depends on at least four factors: the size of the population modelled, the model used, the level of malaria transmission, and the length of the time period simulated; hence creating a model of the run-time of our simulation work-units (each is a model of one scenario) is not quite straightforward.

Hopefully we can come up with a better estimate of work-unit run-time today to at least fix future work sent out, and correct the wrong credit already granted sometime this week.

As always, much thanks for your understanding.
Diggory & the OpenMalaria team

But also make sure you give extra credit to the ones that gave 1/10th the credit due as well.

It seems to be TWO problems :

1st: credit granted multiply per 10 easily observable.

2nd: the "duration correction" of the apply pass from 1.08 to 100.08 !!!
This could explain that many PC's run malaria control in high priority.

Hope this will help
____________

But also make sure you give extra credit to the ones that gave 1/10th the credit due as well.

Hello,

I hope you will correct this problem because I do not receive additional credit ... so it's not just for me and others ...

I stopped BOINC and manually edited client_state.xml to reduce the duration correction factor to 1 for MCDN. All applications were playing well with that (including openMalariaA v6.51 tasks with new WUs which appear to have correctly scaled variable values for ) until I was sent reissues from older workunits.

My quad ran a few of those overnight. Its DCF exploded back up to 100 and is now on a slow decline back to 1 (1% reduction per completion). BOINC is back to immediately running every downloaded task in high priority mode because the DCF gives them an excessive estimated run time.

My laptop hasn't run any reissued tasks yet, but it does have 2 of them queued. They show estimated run times of 0 and 1 second with DCF=1.1145. I'm going to stop BOINC and scale up by a factor of 3600 for those tasks before they start running.

Edit: made that change and the estimates for the reissued tasks are now much more reasonable at 0:56:48 and 1:17:56.
____________
"The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer

My huge credits got back to normal levels, but the expected time to completion is still wrong. I just got a WU with task number 63266432, which originally expected to complete itself after 53000 hours, which makes approximately 6 years.

The estimate is currently going down at a rate of 40 hours per second and now in 18 minutes it has finished 37% of the task.

My boinc manager is pumping out Malaria WUs like crazy, as they all high priority due to the rather exaggerated.

Cheers,

Mr Cheerful

I loose a lot of finished workunits ( not that monster Credit Wu's ) that are marked as invalid. This Wu's have no problems running on my computers and i am really pissed off.
Sry, but in short time today i will switch back all maschines to another projects.
Really inadmissable!

Assuming that most people here are grownups: Does anybody REALLY care about credit?

Workunits arrive on my PC as 87.600:00:00 *HOURS* WUs, but luckely completing in a much shorter time.....
____________

Assuming that most people here are grownups: Does anybody REALLY care about credit?

Dear all,

Yes, we definitely have noticed the huge increases in credit given out over the weekend. We plan to fix the validator/credit granting system soon and correct recently given credit numbers (sorry, we can't really let you keep those extra credits; the graphs linked above should be more than enough explanation)!

As to why, Thyme Lawn seems to have spotted the reason. We're aiming to migrate away from the assumption that all work-units are the same length (they're not, and this has frequently led to a factor 2 or so misestimate of expected work-unit run-time). Unfortunately though, it appears our first attempt at this has gone wrong. Work-unit run-time depends on at least four factors: the size of the population modelled, the model used, the level of malaria transmission, and the length of the time period simulated; hence creating a model of the run-time of our simulation work-units (each is a model of one scenario) is not quite straightforward.

Hopefully we can come up with a better estimate of work-unit run-time today to at least fix future work sent out, and correct the wrong credit already granted sometime this week.

As always, much thanks for your understanding.
Diggory & the OpenMalaria team

Dear all,

Yes, we definitely have noticed the huge increases in credit given out over the weekend. We plan to fix the validator/credit granting system soon and correct recently given credit numbers (sorry, we can't really let you keep those extra credits; the graphs linked above should be more than enough explanation)!

As to why, Thyme Lawn seems to have spotted the reason. We're aiming to migrate away from the assumption that all work-units are the same length (they're not, and this has frequently led to a factor 2 or so misestimate of expected work-unit run-time). Unfortunately though, it appears our first attempt at this has gone wrong. Work-unit run-time depends on at least four factors: the size of the population modelled, the model used, the level of malaria transmission, and the length of the time period simulated; hence creating a model of the run-time of our simulation work-units (each is a model of one scenario) is not quite straightforward.

Hopefully we can come up with a better estimate of work-unit run-time today to at least fix future work sent out, and correct the wrong credit already granted sometime this week.

As always, much thanks for your understanding.
Diggory & the OpenMalaria team

G'Day MalariaControl,
I have been a member of this project for a year but only yesterday attached some computers.
The first thing I noticed was the variation in Work Unit run times, this in itself is not a problem.
However the way credit is granted is of interest to me.
On a WU that runs for 2,600 seconds I might get anywhere from 6 credits to 115 credits, but usually less than 30 credits (which is ordinary) .
On a WU that runs for 3,000 or 4,000 or over 5,000 seconds I would expect to get much more credit than on the shorter work units (some are as short as 450 seconds and a few credits).
This is not the case as can be seen with one example of a 4,780 second WU getting just 1.97 credits (to see this WU my computers are not hidden, it is the X6 Linux machine).

Can this be explained to me please ??

(P.S. You may of noticed the large credit increases that some have been awarded and I noticed even before attaching, but have you noticed the just as large credit decreases?)

Thanks
Conan

Can this be explained to me please ??

Posted with some foreboding, as i am aware of the work level this kind of thing causes .... below may help, if so great, if not cann it.

I was idly nosing around seeing the effects of the fix attempts, and came across a possible anomaly - could be I saw it mid way been fix activities ... dont know. Conan and I both have a 1090T, and over the last 12 hours or so been running the same WUs (edit: WU type), however the credit awards are drastically different. The drastically different credit levels on each of our avatars brought it to my attention.

I appear to be getting - for the most part - a substantial fixed credit award of 4,472 for each WU (a small number of exceptions), he is getting awards back to broadly speaking what it used to be. The only difference would appear to be I am running Win 7 64bit, he is running Linux.

Purely an observation for the melting pot, I'll creep quietly back into my box :)

I echo Thyme's thoughts re the credits, hope it settles for you soon.

Regards
Zy

Assuming that most people here are grownups: Does anybody REALLY care about credit?

sry, i loose about 120000 points with 5-6 Computers are worket on, you can see in my statistic. I got 3 Wu's with a higher Creditvalue and they are not invalid.
I want my ~120000 credits BACK!

Hi All,

just want to add some results of my Win7 64bit machine.

_____________WU_________SECONDS_________CREDIT
________24940136________12,874.28__________79.30
________25051611_________1,211.38___________1.35
________25051536_________1,967.33_______4,116.35
________24937809________18,917.16_________145.61


I joined yesterday and with this type of credit rules I will quit the project after i ran all the outstanding jobs.
Maybe I come back when i see a message on the news that this is truly fixed.

Hi All,

just want to add some results of my Win7 64bit machine.

_____________WU__________SECONDS__________CREDIT
________24940136________12,874.28__________79.30
________25051611_________1,211.38___________1.35
________25051536_________1,967.33_______4,116.35
________24937809________18,917.16_________145.61


I joined yesterday and with this type of credit rules I will quit the project after i ran all the outstanding jobs.
Maybe I come back when i see a message on the news that this is truly fixed.

what about the already granted ones,
stats are all wrong everywhere now,

wtf are u doing with my credits in malaria? Are complete overfreaking now?
You canceled even today over 70k credits!!!!!!!!!!!!!!!!
Is that how you handle my work ( with 6 Computers ) for YOUR project?
Thx, but for now thats it, finally. Good for your when you say, we can ask nobody for staying in, thats how we handle problems.
"if a few people decide to quit malariacontrol.net over this issue, I guess we can't expect much less".
I see pics can even fly.
OVER and out!
PS
Was it really a issue or more even planed to get more cruncher "invited" in this project?
How much is the difference before/after????
____________

wtf are u doing with my credits in malaria? Are complete overfreaking now?
You canceled even today over 70k credits!!!!!!!!!!!!!!!!
Is that how you handle my work ( with 6 Computers ) for YOUR project?
Thx, but for now thats it, finally. Good for your when you say, we can keep nobody for staying in effect of our handling problems.
"if a few people decide to quit malariacontrol.net over this issue, I guess we can't expect much less".
I see pics can even fly.
OVER and out!
PS
Was it really a issue or more even planed to get more cruncher "invited" in this project?
How much is the difference before/after????

Just read this thread for an explanation of other crunchers reaction and what the Project Admins are doing about it. It also covers the topic of how long the adjustments to total credit may take, and that no changes to RAC are proposed as this will self correct over about a month.
____________
Go away, I was asleep

Said a Russell, 3 Shih-Tzus & a Bischeon Frize

Before this issue started I had 108,767 credits as of November 20th. I had a hand full of WU's with outlandish credits and agree those should not stand and should be backed out. The problem is, now I'm down to 4,546 credits. Far too much has been backed out. I had 80,488 credits a month ago on October 25th and even a year go on October 28th 2009 I had 51,621 credits. I don't know if you are aware of this over correction, and wanted to bring it to your attention.

Krunchin

I think the over correction on credit (removed the artificially high credit given and most of the credit earned over the last 2 - 3years - for me) is currently being evaluated and corrected.

I can quote my credit before the current removal (925,989) from my BOINCStats report in my signature. But, my loss is far behind yours (2.5million) and Mikey's (4.5million or near enough).

The Admins (and you, as a Mod, know better) say they will restore it all in the next few days.

I am crunching quite vigorously now, and over the next few days. So, the output I make now will need to be added to the total I quote, and the same for you and the rest of us.
____________
Go away, I was asleep

Said a Russell, 3 Shih-Tzus & a Bischeon Frize

Strike, seems like a good option to this mess!
I join shortly...To many cpu at work. this will take some time.

Almost my every computer is in strike here in malaria:

ID: 82543 AuthenticAMD
AMD Sempron(TM) 2600+ [Family 6 Model 8 Stepping 1]
(1 processors)

ID: 82543 AuthenticAMD
AMD Athlon(tm) 64 X2 Dual Core Processor 6400+ [Family 15 Model 67 Stepping 3]
(2 processors)

ID: 123225 AuthenticAMD
AMD Phenom(tm) 8650 Triple-Core Processor [Family 16 Model 2 Stepping 3]
(3 processors)

ID: 139998 AuthenticAMD
AMD Athlon(tm) II X2 250 Processor [Family 16 Model 6 Stepping 2]
(2 processors)

ID: 121780 uthenticAMD
AMD Phenom(tm) 8650 Triple-Core Processor [x86 Family 16 Model 2 Stepping 3]
(3 processors)

ID: 29505 AuthenticAMD
AMD Athlon(tm) 64 X2 Dual Core Processor 5200+ [Family 15 Model 67 Stepping 2]
(2 processors)

ID: 172735 thenticAMD
AMD Athlon(tm) II X4 640 Processor [Family 16 Model 5 Stepping 3]
(4 processors)

ID: 121431 AuthenticAMD
AMD Phenom(tm) II X4 945 Processor [Family 16 Model 4 Stepping 3]
(4 processors)

ID: 172861 AuthenticAMD
AMD Athlon(tm) 64 X2 Dual Core Processor 5600+ [Family 15 Model 67 Stepping 3]
(2 processors)

ID: 173007 AuthenticAMD
AMD Athlon(tm) II X4 640 Processor [Family 16 Model 5 Stepping 3]
(4 processors)

ID: 125416 AuthenticAMD
AMD Athlon(tm) 64 X2 Dual Core Processor 5600+ [Family 15 Model 107 Stepping 2]
(2 processors)

ID: 125298 GenuineIntel
Intel(R) Core(TM)2 CPU 6420 @ 2.13GHz [x86 Family 6 Model 15 Stepping 6]
(2 processors)

ID: 172808 AuthenticAMD
AMD Athlon(tm) II X4 640 Processor [Family 16 Model 5 Stepping 3]
(4 processors)

ID: 173078 AuthenticAMD
AMD Athlon(tm) II X4 640 Processor [Family 16 Model 5 Stepping 3]
(4 processors

ID: 100677 AuthenticAMD
AMD Athlon(tm) II X3 445 Processor [Family 16 Model 5 Stepping 3]
(3 processors)

As it is the WUs still keep coming with abnormal ETC (estimated time of completion), thereby screwing up the priorities for the queues for the various projects I have running on my PC. During the days of credit-craze this was responsible for 6 of my PCs devoting all their time on Malaria, with as resultant extra credit output, which now has been slammed back to the level of about a week ago. While I can live with that, I want Malaria to have their WUs send as WUs of 1.5 hour -or as long as it takes for any of my PCs- and not as 76.800:00:00 hour WUs -as I receive them now-.
____________

While I can live with that, I want Malaria to have their WUs send as WUs of 1.5 hour -or as long as it takes for any of my PCs- and not as 76.800:00:00 hour WUs -as I receive them now-.

• 
  
• stop BOINC.
  
• edit the file client_state.xml in your BOINC data directory using a plain text editor.
  
• search for the string malariacontrol. That should find the line https://malariacontrol.net/.
  
• search for the string duration.
  
• change the value between the XML tags to 1.000000. The line will end up as 1.000000.
  
• save the file.
  
• restart BOINC.
  

Without that change your DCF will very gradually reduce to a reasonable level on its own (by 1% every time you complete a task).

Setting DCF to 1 might still be too high for your computer. If you find it is just repeat the process and reduce the value further (from the durations you gave I wouldn't be surprised if you had to set DCF as low as 0.002000).[/b]
____________
"The ultimate test of a moral society is the kind of world that it leaves to its children." - Dietrich Bonhoeffer

Dear Project managers, developers and moderators

I Propose you :
1st) to put the whole project in "test mode" then only peoples who chose the risk, will continue to calculate.

2nd) explain how to change the value of in "client_state.xml" and ""client_state_prev.xml" from 100.xxxxxx to 1.xxxxxxx
for who are choosing running tests appli and then stop to run only in high priority.

3rd) communicate this dispositions on front page.

Be sure to my consideration, I 'me with you and be confidant that you will correct this ... holly mess.

Sincerely
OlaV+++
____________

Ok, i thing i do calm down now. Sry for my hard words before.
I will join again, after the creditsystem working properly.
It's ok when the very high graded Wu's are resetet to a normal level.
But this Wu's need to be valued too.
____________

I haven't stopped, but am only getting about 1/100th the credit I was before.

Ouch...

Ditto
____________
Go away, I was asleep

Said a Russell, 3 Shih-Tzus & a Bischeon Frize

I haven't stopped, but am only getting about 1/100th the credit I was before.

Ouch...

Ditto

Well a new low awarded credit amount, 0.26, for a WU over 2500 seconds.

I have set NNW on my Windows machines due to all the errors they are running.

If the credit situation has been fixed then what is a valid amount for a Work Unit?

I don't believe that I am getting much of a return on my input.

My Linux machines are still getting ultra low credit amounts.
My Windows machine only had a few thousand credits and they are nearly all gone with errors for all other work.
My Host totals don't show the amount of awarded credit.
My Total does not show my awarded credit.

As I did not receive any of the huge credit amounts for work units but I did have 4 that that were between 1700 and 4000, so these are also classed as huge amounts?

I only started crunching here a few days ago so I do not know what is termed a valid amount of credit for a returned WU, so can someone enlighten me please?

But from my stand point I don't see that credit has been 'Fixed', either as a 'fixed' amount or 'fixed' the problem.
The currently returned work units are still giving low credits.

Thanks for your time
Conan
____________

I thinking about totally ban my cpu work, saves my electricity bill, and I am still be able to do work with gpus, now 1 million+/day, next year target will be 2 million/day. Now back in 101 place by total credit in world and going to top 100 soon. 20 000 credit/day in malaria does not matter, no differences to my rac. The only thing that make a differences is what cause you want help, we need more credits here! Please make a gpu project available here in malaria.

But from my stand point I don't see that credit has been 'Fixed', either as a 'fixed' amount or 'fixed' the problem.
The currently returned work units are still giving low credits.

oM-B is for my oppinion to low valued with credits what i see in the last Day's.
And what about my and other volunteers lost Credits?
I guess i'm missing more than 200k points, genrated with a huge "maschinepark".
Ok Leute. Mal ehrlich, habt ihr eigentlich mal intensiv eurer Kreditsystem vorher überprüft?
Was tut ihr gerade um die Probleme zu beseitigen, auch die gelöschten WU's könnt ihr nicht einfach so in Luft auflösen.
____________

But from my stand point I don't see that credit has been 'Fixed', either as a 'fixed' amount or 'fixed' the problem.
The currently returned work units are still giving low credits.

The correction has only been made for host, user and team total credits. Neither RAC nor the credit shown for individual validated tasks have been touched. My guess is that tasks weren't adjusted because they're normally deleted within a few days of being validated.

The validator for openMalariaA tasks has been disabled since the project team started the credit correction measures which Nick outlined in his news post and won't be re-enabled until Monday.

openMalariaB tasks weren't affected by the credit problem and are still being validated and credited (as indicated on the server status page).

As others have said the credits noe are grossly undervalued. https://malariacontrol.net/results.php?hostid=143895

My latest task has 87600 hrs to completion! 10yrs if I last that long.

Well a new low awarded credit amount, 0.26, for a WU over 2500 seconds.
If the credit situation has been fixed then what is a valid amount for a Work Unit?
I only started crunching here a few days ago so I do not know what is termed a valid amount of credit for a returned WU, so can someone enlighten me please?
Thanks for your time
Conan

But from my stand point I don't see that credit has been 'Fixed', either as a 'fixed' amount or 'fixed' the problem.
The currently returned work units are still giving low credits.

The correction has only been made for host, user and team total credits. Neither RAC nor the credit shown for individual validated tasks have been touched. My guess is that tasks weren't adjusted because they're normally deleted within a few days of being validated.

This might be true, but there is no way to tell how much credit has been granted (?)

Was credit granted for all the already 'pending' / 'completed waiting validation', ie, completed days ago or are those stiil to be credited when the validator comes back on ?

I find this all very confusing and the limited/short explanations not all that helpful to us volunteers, at least me anyway.

My credit has been restored to a number close to what it was, but since I don't know the exact total before all this happened i can only guess, it seems acceptable.

It is very hard to tell how much was granted for all the previously miscredited ones.

I need a break from all the stress caused, thus i'm taking a break from the boinc world for an indefinate period, see everybody sometime in the future.

Hi

We took the users' total credits from our backup on Friday the 19. of November (Before all went wrong).
Then we have listed all the valid results from the 19. of November to last Friday the 26. of November (After having stopped the boinc server).
Using some metrics in the results database table (flops estimate, cpu time) we have then calculated a fair (we have checked a large sample of results!) credit per job for all the users.
This means that the high amounts and the low amounts have been corrected as well.

Nick did already send a news about this on Friday.

Thanks
Guillaume

Thanks Hardy,

No credit has been added to my account.
From what I can see my totals have not moved at all, my Linux and Windows machines have the same credit they had a few days ago.
Unfortunately for me all those low credit work units have now been purged so I can't see what was granted to those work units.

I had a lot of the very low credit types (as low as 0.26 credits for a 2,500 second work unit).
No evidence anything has changed.

But thanks for your effort in trying to work out the problems, much appreciated.

Conan
____________

Thanks Hardy,

No credit has been added to my account.
From what I can see my totals have not moved at all, my Linux and Windows machines have the same credit they had a few days ago.
Unfortunately for me all those low credit work units have now been purged so I can't see what was granted to those work units.

I had a lot of the very low credit types (as low as 0.26 credits for a 2,500 second work unit).
No evidence anything has changed.

But thanks for your effort in trying to work out the problems, much appreciated.

Conan
____________

Same here Conan, nothing chainged yet. But i can wait. I just hope @ the end my allover credits back on ~200k+. ;)
____________

Same here Conan, nothing chainged yet. But i can wait. I just hope @ the end my allover credits back on ~200k+. ;)

Thanks Hardy,

No credit has been added to my account.
From what I can see my totals have not moved at all, my Linux and Windows machines have the same credit they had a few days ago.
Unfortunately for me all those low credit work units have now been purged so I can't see what was granted to those work units.

I had a lot of the very low credit types (as low as 0.26 credits for a 2,500 second work unit).
No evidence anything has changed.

But thanks for your effort in trying to work out the problems, much appreciated.

Conan

Hi Conan

All work units from last week should have had credit re-granted on the basis of how long they were run, so low credit work-units should have been pulled up to. These were however granted directly to hosts/users/teams and not to the work-units in the database, so looking at the individual work-units wouldn't have shown the updated credit anyway.

Avoiding granting excessively low credit is really the main reason the validator has been left off so long, and we hope to have it back online soon.

Diggory

Hi, Am I still receiving anomalous credits for the work performed?

https://malariacontrol.net/results.php?userid=35400&offset=0&show_names=0&state=3

Disregarding the ver 6.49 WU, the ver 6.52 credits seem inconsistent for the CPU times involved. I'm seeing credit rates from 11 credits per hour to 168 credits per hour. This doesn't seem right.

Rob

Hardy said: No, credit per hour was never all that high on openmalaria, but the new BOINC credit system should make the credit granted by all projects comparable.

Well, it's not 20,000+ credits, but this still seems too high:

Workunit 25112488

name wu_968_158_259577_0_1290823467
application openMalaria: A simulator of malaria epidemology and control (Branch A)
created 27 Nov 2010 2:04:27 UTC
canonical result 63426382
granted credit 315.20
minimum quorum 2
initial replication 2
max # of error/total/success tasks 4, 20, 10

27 Nov 2010 2:18:22 UTC 27 Nov 2010 4:18:44 UTC Completed and validated 1,989.16 1,972.86 315.20
4 27 Nov 2010 4:17:18 UTC 28 Nov 2010 4:34:11 UTC Completed and validated 2,566.53 2,537.25 315.20

Credits are going awry again as I have just recieved https://malariacontrol.net/workunit.php?wuid=25143283[/url]

Bobcat, you didn't include the elapsed time or cpu time (credit granting is based on these).

Hi.

Since your talking about low credits, this is since you started the validator

back up.

https://malariacontrol.net/workunit.php?wuid=25154027

28 Nov 2010 4:13:34 UTC__Completed and validated__3,996.94__3,996.94__4.69

____________