• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

We still having stats problems?

T

Tobit

[H]ard|DCer of the Month - March 2010/May 2011
Joined
Dec 12, 2005
Messages
5,145
Anyone else have issues with stats at Stanford? It says my last WU was turned in @ 06:55:22 but I've been folding like a MOFO all day turning in at least 1 x GPU2 WU every 1.5 hours since then and nothing has changed. Should have seen some kind of update on Stanfords site I would have thought. Someone else has noticed some 'issues' and posted something on FF. Weird.

 
As I mentioned briefly in this thread, I'm experiencing a major decline in my production lately. I received a slew of P3062 WUs and almost very one underwent an EUE. I know that's part of the problem but I still think there's another issue, possibly unaccounted WUs at Stanford's end. I has happened before.

 
Yea, I'm having a few issues too. PS3 units uploading like clockwork every 3 hours but a half day of goose eggs on [H]ardFolding and EOC. Plus a box on notfred's cd can't upload.

Come on Stanford, the Chimp Challenge kicks off in 10 days, we NEED the stats working by then. :eek:

 
Stats on my page at Stanford are slowly trickling in finally. Although, it suddenly thinks I've added a new CPU as it now has me at 4 instead of the 3 I've been running on all along. Weird.

 
Tobit said:
Stats on my page at Stanford are slowly trickling in finally. Although, it suddenly thinks I've added a new CPU as it now has me at 4 instead of the 3 I've been running on all along. Weird.

Click to expand...

They've got me listed at 7/4 rather than the 6/4 I've been at for a while. I would suspect a borg finally checking in, but the points tell me otherwise.

 
^ ^ Stanford never had the precise number of processors in my farm either. I don't know exactly how they go about estimating this figure.
 
APOLLO said:
^ ^ Stanford never had the precise number of processors in my farm either. I don't know exactly how they go about estimating this figure.
Click to expand...

It's supposed to be per client based on the unique machine id#. If you reinstall it can create a new number and if you clone, multiple boxes can have the same number. Its usually pretty accurate based on what I remember.

 
I have not had any EUE's .... well i did but I fixed it.

I go three EUE's on two boxen.
I thought they were 100% stable, been folding for months just fine.
Got two EUE's ....

I went a raised the vcore of those boxen, one notch up, and poof no more EUE's

These new work units are just more CPU and RAM intensive.

Give it a try on your side and let us know.

I have not had an EUE on any of my 7 boxen since I did this.!


 
^ The EUEs on my end are only with P3062 WUs, and they happened on systems that have been stable since last year. Moreover, the systems affected are of widely varying architectures. If the problem expands to include other WUs in the future I will definitely lower the OC, but the instability of the 3062s are well known. They are this year's equivalent of the P2652, which was notorious last year for EUEs.
 
Same Problems here, I didn't raised the vcore but lower the oc but 200 mhz and still doing the same thing. Finally had to delete the work unit, 2 other boxen had to delete WUs twice :eek:
 
Considering I'm at 28k per day and I showed up 8k yestereday... yeah something is off... between that and lost WU's this is getting a bit discerning if the science we are churning is getting lost in never never land or if it is being used. This week has been the worst of the nearly 3 months riddled with server problems, out of work, stat problems, etc. I just hope they fix it soon!

One thought is if credit is not given for WU's we chugged through do they at least show up on the servers and not get reissued or are they being reissued and thus wasting resources? I have a feeling I'm going to need to make a post on the forums at F@H... I pray its not wasting resources...

 
Sunin said:
Considering I'm at 28k per day and I showed up 8k yestereday... yeah something is off... between that and lost WU's this is getting a bit discerning if the science we are churning is getting lost in never never land or if it is being used. This week has been the worst of the nearly 3 months riddled with server problems, out of work, stat problems, etc. I just hope they fix it soon!

One thought is if credit is not given for WU's we chugged through do they at least show up on the servers and not get reissued or are they being reissued and thus wasting resources? I have a feeling I'm going to need to make a post on the forums at F@H... I pray its not wasting resources...

Click to expand...

+1

I like my points just as much as the next guy..... But I hope the WU are finding their way home even if we are not getting the points.

I'm down to 7k ppd and FAHMon shows that even with all the nasty WU I have I should be at 15.5k

 
Yeah, I think there are problems again. There's no way I should have 0 points today with a PS3 running.


 
I think the stats from stanford have been tanked for a couple of days now, but they're complete out of order now. Its not looking good right before the challenge, and I sure hope things get fixed and are stable before the 5th.

 
Axdrenalin said:
I think the stats from stanford have been tanked for a couple of days now, but they're complete out of order now. Its not looking good right before the challenge, and I sure hope things get fixed and are stable before the 5th.
Click to expand...
It doesn't look good but the challenge is still more than a week away. Maybe someone should inquire if there's an ongoing stats problem at the Folding forums?
 
I can see the stats on 3rd party sites such as EOC and [H] having issues when Stanford was generating empty reports for awhile. But, when stats are missing from my page @ Stanford, something is wrong. Some of them have been updated but it still looks like I am short 10 WU's. Oh well, trying to be patient but, as a newbie, I am anxious to see my numbers climb!

 
Vijay finally posted some info about this. See his post in this thread here:

http://foldingforum.org/viewtopic.php?f=18&t=2300

Basically, all the really small WUs going out lately are overloading the stats servers. These are the WUs that P4's crank out 10 to 12 of a day along with the small GPU2 WUs. He indicates he is taking drastic measures today to fix things. :eek:

 
That's true, especially when we know the standard console client make up 80% of the overall production (not everyone is as good as us at handling SMP client issues). Multiply this with the amount of uploads/downloads daily and it's like a DDoS attack on the servers.

 
Tobit said:
Vijay finally posted some info about this. See his post in this thread here:

http://foldingforum.org/viewtopic.php?f=18&t=2300

Basically, all the really small WUs going out lately are overloading the stats servers. These are the WUs that P4's crank out 10 to 12 of a day along with the small GPU2 WUs. He indicates he is taking drastic measures today to fix things. :eek:

Click to expand...

So, are we saying that even though out data is getting returned to stanford successfully, the smaller packets are overwhelming the statistic generating servers and crashing them? I know I've had tons of WU's sent up, but I'm not seeing any results (even on the stanford site) that they are being processed / accounted for.


 
Well, as long as the WU's are getting sent in and are being used to further research I don't have a problem with the delayed stats. I can't speak for the other users out there however. It could turn some people away from folding. :(

 
That makes sense to me... I turned in something like 96 wu's a few days ago.... and today I have already turned in 14 for a whopping grand total of 350 points:(

These smaller wu's don't really bother me... as I've said before, if it's work that needs to be done, so be it.... but if they are borking the stat's servers, perhaps I should just kill my borg until they get things fixed?

I don't know.... it's frustrating...... I was really hoping I would be seeing some of the "bonus point" wu's for the classic clients running -advmethods... all of my borged boxen are and I don't think I've received any yet.....

Hopefully all this is cleared up before the challenge begins.....

 
rmdashrrootsplat said:
Well, as long as the WU's are getting sent in and are being used to further research I don't have a problem with the delayed stats. I can't speak for the other users out there however. It could turn some people away from folding. :(

Click to expand...

I'm one of the people that think about the pernts thing about like you, but, like you, I also think it would be a "turn off" for some of the "heavy duty" folders.

Hell, they've invested mucho denaro (a lot of $) in some unbelievably nice boxen. Of course the cause is most commendable, but I think it's only fair that anybody (especially the heavy folders) should see some immediate results (ie pernts).

I mean, good God :eek:, I wish I had a nickel for every time I've read where a folder has to prove to some moron by showing him/her some written results from folding. (proof that folding helps cure diseases) :rolleyes:

When it's all said and done I just wish that Stanford would get offa' their behinds or whatever and fix this "no credit for WU's thing" before the Chimp Challenge. (only an echo, I read that in digital_exhaust's post above :))

Don't nobody get their shorts in a twist. This is only my opinion, ya' know, that ratty ole' 1st amendment :p)

Edit: In no way am I putting the terrific staff of F@H down. I've read where they're aware of the problem and it's caused by some early return WU's overloading the servers. I'm just following a thing I learned many years ago :"I ain't believin' nuttin' til' I see's it".


 
From what I have read and understand, the WU we have completed are being succesfully captured. The work will not have been in vain. That's a good thing.

The points, as everyone has stated, is secondary (albeit frustrating not to see it, as human nature requires immediate gratification). They points aren't lost, but will over time (hopefully not too long), will catch up. Those that are missing WU will eventually find them on their stat sheets. It does look as if they are trying to at least once a day get the points up. Kudos for getting that much.

Now, if the stats servers are getting borked, maybe we need to donate to them a few servers to keep up :)

 
digital_exhaust said:
I don't know.... it's frustrating...... I was really hoping I would be seeing some of the "bonus point" wu's for the classic clients running -advmethods... all of my borged boxen are and I don't think I've received any yet.....
Click to expand...
Has Stanford started to release those bonus WUs yet?
 
APOLLO said:
Has Stanford started to release those bonus WUs yet?
Click to expand...

They are part of the promlem.
Add in the small size of the GPU2 work units and the servers are getting hammered.

My understanding of it is ..............
The work servers are just ok ..... :)
Where its breaking down is getting the data from the work servers to the stats servers.
What was taking 1 hour to run is now taking upto 5 hours due to the amount of data being passed across.

Luck .......... :D
 
APOLLO said:
You mean the new tiny WUs are the bonus WUs mentioned in this thread: http://www.hardforum.com/showthread.php?t=1299458
Click to expand...

As I understand it , Yes.

From Vijay Pande over on the Support forum ..............
This is precisely the problem. I had hoped that recent fixes there would have done the trick. Since it hasn't, I'm going to take more strong action today (potentially stopping new assigns there) until this can get under control.

The good news is that the stats are all there, it's just taking 5x longer to get them inputted. Once we get those new projects back to the right size, this should go back to normal.

Luck .............. :D
 
Well.... I hope they get the points fixed soon.... because if the 45XX wu's are indeed the "bonus point" wu's... the ones that we are supposed to be getting x1.5 credit for...something is seriously wrong.

I've been getting a ton of these and I have yet to see over 10 points for them... unless I'm misunderstanding something here (which I probably am:p) that doesn't really seem like a "bonus" to me... other than I'm seeing 5 minute frame times.......

 
digital_exhaust said:
Well.... I hope they get the points fixed soon.... because if the 45XX wu's are indeed the "bonus point" wu's... the ones that we are supposed to be getting x1.5 credit for...something is seriously wrong.

I've been getting a ton of these and I have yet to see over 10 points for them... unless I'm misunderstanding something here (which I probably am:p) that doesn't really seem like a "bonus" to me... other than I'm seeing 5 minute frame times.......
Click to expand...
I've seen a 15 pt WU as well. Stanford may be implying that the amount of time these WUs take to process, at the end of the day we're left with a 50% increase in points versus standard units. However, I could be wrong and these aren't the 'bonus' WUs Stanford is talking about.
 
APOLLO said:
As I mentioned briefly in this thread, I'm experiencing a major decline in my production lately. I received a slew of P3062 WUs and almost very one underwent an EUE. I know that's part of the problem but I still think there's another issue, possibly unaccounted WUs at Stanford's end. I has happened before.

Click to expand...

The 3062’s are more chewy but if you have had a bunch of early ends something is really wrong.

Might be heat or an over clock set a tad high but the WU as a whole is very stable. It’s also one of the reasons everyone’s points took a hit. In other words, it sucks ;)

As for Stanford, well, for as much as many of us have invested in high end computers with our personal farms I find it awfully hard to believe they can’t get their hands on a few new servers and cleanup the software they use to distribute work.

To Stanford’s credit I have yet to have them actually lose a WU. They may get lost for a while but somehow somewhere they do get found.;)


 
BillR said:
The 3062’s are more chewy but if you have had a bunch of early ends something is really wrong.

Might be heat or an over clock set a tad high but the WU as a whole is very stable. It’s also one of the reasons everyone’s points took a hit. In other words, it sucks ;)
Click to expand...
MY OC are stable and I live north of the border so our temperatures are still cool here. It's definitely the WUs.

On my dual quads I have more than 2 SMP clients running, yet if one of them is a 3062 there's a high probability it will crap out. If it's a hardware issue, then why don't the other WUs on the same machine running concurrently also suffer an EUE? All of my quads are clocked at 2.8GHz or slower, and we all know that is way too low a frequency for the core architecture to raise doubts about hardware stability. The fact only one specific WU is problematic, and the fact others have also experienced the same issues points to an errant WU.
 
Nope....not here anyhow....

And now the stats site at Stanford is down.... apparently for 24 hours:(

Hopefully they get things fixed soon...

 
alan2308 said:
Some people are getting an update here and there, all at different times.
Click to expand...

I'm still not sure it was a full update for all the WU I've done.

They better add more boxen to deal with all of this production.

 
Kendrak said:
I just got an update on the stats..... anyone else?
Click to expand...
I did as well but it was an insubstantial amount, approximately 8K. Hopefully by the beginning of the work week all will be normalized.

 
I got a small update of 4000 pts, which is not what I expected since I witnessed 5 SMP WU uploads including a 3065 worth 2144 points this afternoon.

 
You must log in or register to reply here.
Back
Top