End of Beta cores?

WhitehawkEQ · Dec 4, 2012

V7 will not run GPU WU's on the beta core 2.22 anymore, here is what I get:

Code:

22:01:00:WU00:FS01:Downloading core from http://www.stanford.edu/~pande/Win32/x86/NVIDIA/Fermi/beta/Core_15.fah
22:01:00:WU00:FS01:Connecting to www.stanford.edu:80
22:01:00:WU00:FS01:FahCore 15: Downloading 1.88MiB
22:01:05:WU00:FS01:FahCore 15: Download complete
22:01:05:WU00:FS01:Valid core signature
22:01:05:WARNING:WU00:FS01:FahCore has not changed since last download, aborting core update
22:01:05:WU00:FS01:Starting
22:01:05:WU00:FS01:Running FahCore: "C:\Program Files\FAHClient/FAHCoreWrapper.exe" "C:/Documents and Settings/Rodney/Application Data/FAHClient/cores/www.stanford.edu/~pande/Win32/x86/NVIDIA/Fermi/beta/Core_15.fah/FahCore_15.exe" -dir 00 -suffix 01 -version 701 -checkpoint 15 -gpu 0
22:01:05:WU00:FS01:Started FahCore on PID 1320
22:01:05:Started thread 176 on PID 2272
22:01:05:WU00:FS01:Core PID:2212
22:01:05:WU00:FS01:FahCore 0x15 started
22:01:05:WU00:FS01:FahCore returned: CORE_OUTDATED (110 = 0x6e)

It repeats this over and over. This is on 2 GPU's so far, any ideas on what is going on? :shrug:

dfonda · Dec 4, 2012

I gave up on the v7.(Due to these new issues)The last thing I needed was more headaches. The v6 client is running like a champ again.

WhitehawkEQ · Dec 4, 2012

It's not the V7 that is the problem, it's the GPU Beta core 2.22.

dfonda said:
I gave up on the v7.(Due to these new issues)The last thing I needed was more headaches. The v6 client is running like a champ again.

Edward2 · Dec 4, 2012

I haven't gotten any 8057 WU's since early this morning, but I've heard that the 2.22 core is no longer valid. You have to use the 2.25 core, which means that the PPD is dropping by a lot. Maybe we are starting to see where the GPU QRB WU's are going to be PPD-wise.

LuvToGame · Dec 4, 2012

Ok. What exactly is happening? Is core 2.22 now a no-no? Is beta 8057 gone? Or has the server directory structure been changed to eliminate the /beta core folder? Can we still fold 8018's on the 2.22 core?

Please does anyone know?

HayesK · Dec 5, 2012

WhitehawkEQ said:

V7 will not run GPU WU's on the beta core 2.22 anymore, here is what I get:

Code:

22:01:00:WU00:FS01:Downloading core from http://www.stanford.edu/~pande/Win32/x86/NVIDIA/Fermi/beta/Core_15.fah
22:01:00:WU00:FS01:Connecting to www.stanford.edu:80
22:01:00:WU00:FS01:FahCore 15: Downloading 1.88MiB
22:01:05:WU00:FS01:FahCore 15: Download complete
22:01:05:WU00:FS01:Valid core signature
22:01:05:WARNING:WU00:FS01:FahCore has not changed since last download, aborting core update
22:01:05:WU00:FS01:Starting
22:01:05:WU00:FS01:Running FahCore: "C:\Program Files\FAHClient/FAHCoreWrapper.exe" "C:/Documents and Settings/Rodney/Application Data/FAHClient/cores/www.stanford.edu/~pande/Win32/x86/NVIDIA/Fermi/beta/Core_15.fah/FahCore_15.exe" -dir 00 -suffix 01 -version 701 -checkpoint 15 -gpu 0
22:01:05:WU00:FS01:Started FahCore on PID 1320
22:01:05:Started thread 176 on PID 2272
22:01:05:WU00:FS01:Core PID:2212
22:01:05:WU00:FS01:FahCore 0x15 started
22:01:05:WU00:FS01:FahCore returned: CORE_OUTDATED (110 = 0x6e)

It repeats this over and over. This is on 2 GPU's so far, any ideas on what is going on? :shrug:

Need to manually delete the copy of Core_15. Pause the client, then delete the Core_15 folder and contents. The required core should be downloaded when the client is restarted.

The "update" issue most likely rerelated to manually downgrading the core to earlier version than the client had previously downloaded...

The log shows problem with the core_15 copy in the "beta" folder. The non-beta p762x wu, which also require core_15 v2.25, will run the core from the "Fermi" folder, so similar "update" problem will occur if that copy has also been downgraded. v2.22=4576KB. v2.25=7897KB.

HayesK · Dec 5, 2012

LuvToGame said:
Ok. What exactly is happening? Is core 2.22 now a no-no? Is beta 8057 gone? Or has the server directory structure been changed to eliminate the /beta core folder? Can we still fold 8018's on the 2.22 core?

Please does anyone know?

see "Beta Team" topics about p8057 forcing core_15 v2.25 and new GPU projects p8070-8074 being released for beta testing using the unified GPU/SMP benchmarking scheme.
http://foldingforum.org/viewforum.php?f=66

Forum is "read-only" for non-beta team members. Info on Beta Team and how to join.
http://foldingforum.org/viewtopic.php?f=16&t=8

WhitehawkEQ · Dec 5, 2012

Hey HayesK, I think you need a few more folders

<= 7-2600K, 10-i7, 3-C2Q, 10-GPU => folding for OCF T32 as HayesK

LuvToGame · Dec 5, 2012

Ok.

Looks like the weather isn't cool enough down 'ere in the South to GPU fold anyway esp when we get hit with a drop in production of greater than 33% for apparently no good reason except someone's OCD need for uniformity. I'm shutting down the Fermi's until this is resolved and/or the 807x's start shipping out in force.

If anyone else has some information to further this discussion please continue below.

Edward2 · Dec 5, 2012

I've got one of the new 8073 WU's on my GTX580 right now, but HFM.net shows 0 PPD, because the WU does not appear in the psummary or psummaryC lists. I do not know if it has a QRB or not. FAH v7 shows it as 27K PPD. Here is a comparison:

WU 762x = 33K PPD
WU 8054 = 17K PPD
WU 8057 = 200K PPD
WU 8073 = 27K PPD

BrotherJohn · Dec 5, 2012

No QRB on the 807x WU's.

EarthDog · Dec 5, 2012

8054 on 680 @ 1241 core is around 24k.

680 + 3770 + 3570 (both CPUs at 4.5Ghz) look to net me 70k+.

dfonda · Dec 5, 2012

So these are Beta team only? Its great we have so many folks on the team willing to test.

My v7 client must have had some other problem.

LuvToGame · Dec 5, 2012

Now wait a minute. . .

The new 807x's are supposed to be awarded points according to the "new GPU benchmarking scheme." Doesn't that equate to QRB? No?

This just gets better and better.

I could tolerate the loss of recognition in losing the points, but to lose productivity in the same act of losing core 2.22?

LuvToGame · Dec 5, 2012

dfonda said:
So these are Beta team only? Its great we have so many folks on the team willing to test.

My v7 client must have had some other problem.

Yeah, yeah I hear ya. Who would have guessed there were so many closet beta testers out there, huh? :chair:

HayesK · Dec 7, 2012

WhitehawkEQ said:
Hey HayesK, I think you need a few more folders

Actually, the GPU count was low. Updated sig to 16 GPU. (Black Friday deals +3 Kepler, +3 Fermi)

WhitehawkEQ · Dec 12, 2012

Just for grins I added the beta flag to 1 of my SMP slots to see what I get, will let you know if anything comes up

GoD_tattoo · Dec 12, 2012

HayesK said:
Actually, the GPU count was low. Updated sig to 16 GPU. (Black Friday deals +3 Kepler, +3 Fermi)

WhitehawkEQ · Dec 12, 2012

Well, I set the beta flag for the SMP slot and I'm getting 2 WU's that is not on the Project Summery list, 11050 & 11051. Up untill now I have never had these type of WU's

Leonardo · Dec 12, 2012

The 8057 beta project ended. Folders finished the project, which was a test of the new benchmarking, base points, productivity, and Quick Return Bonus scheme. Pande Group has re-benchmarked all active GPU projects and has set new, mostly (or all) higher base point levels. The QRB formula(s) are set. QRB is not yet implemented. PG announced that they will implement QRB for GPU projects after they have worked on issues on their end. They did not specify what the issues are. Speculation is that servers do not yet have the necessary code updates - not an easy task, to facilitate QRB at this time.

We'll just have to be patient. We also need to remember that the 8057 was a test and that real work units may or may not receive such high point awards.

That is all, carry on.

So, at this time, I do not know whether my mid-term Folding strategy will be pure CPU, GPU, or a mix.

End of Beta cores?

Premium Member

Senior Golfer

Premium Member

Member

Member

Member

Member

Premium Member

Member

Member

Member

Gulper Nozzle Co-Owner

Senior Golfer

Member

Member

Member

Premium Member

Member

Premium Member

Member

Similar threads