Anyone else run into this nightmare yet? It has a woeful ppd and really long runtime.
If a workunit is going to take more than 24 hours to run on a 2600K @ 4.5 I want
more than 22240 ppd for it!

I've had them none stop for about 10days now.  I'm not getting much more PPD than you and i'm running a pair of hex core westmeres @2.4

On my 2600K, I've been getting nothing by 609x WUs for weeks.  I noticed that you are running the V7 Client.  I'm running the Console Client and this is one of those areas where I always wonder if the Client is the difference.  I haven't seen a single one of these 7647's that you guys are working.  I know that I'm not receiving A4 WUs on the Console Client, so that may be why.  I see that this WU you are reporting is an A4.  I'm guessing that the 7647 is also getting the "10% bonus points program" that was announced last month by PG?

All A4 WU are under the 10% bonus program.
 
As for different clients that does seem to make a difference in what WU you get, especially under windows.  I've always found that v6 gives a3 WU and v7 gives a4 WU. 
 
These 7647's remind me of the dreaded 6701's from a couple of years back - lots of work for little reward

I haven't gotten one that I seen.  All I have been getting lately are the 8004 & 8049.
 
Is there a program that can track what WU and their PPD value you have done?  I know the GPUTracker did, but the FAH v7 doesn't.

im getting 17 k on a 970 hex at 4.4 i should be getting 38k normally this is getting silly

These P7647 SMP WUs give a lot more points than the same WUs for GPU3 we had a few months ago.

I don't know if anyone still remembers GPU3  P7640/41/42/43/44 WUs, which had some problems. These WU's didn't fully utilize your GPU and also had a lot of fail WUs and Stanford rolled them back to beta:
GPU Projects 764x rolled back to beta!

A few days later Stanford moved some of these GPU3 WUs to SMP: 764x projects going to advanced

So P7647 (and P7645, P7646) SMP WUs are the only WUs I know you can directly compare to GPU3 WUs and I have to say that the points difference is HUGE.

I've searched my logs and found that my GTX460's had almost the same performance as Simba's 2600K:
P7647 -  2600K @   4.5GHz: TPF = 12:48 | PPD = 22.3K | Points = 19.800
P7644 - GTX460 @ 800MHz: TPF = 12:50 | PPD =  8.7K | Points =   7.718

As you can see the same WU gets ~2.5x more points on SMP than on GPU3.. Edit: so that's a high (QRB) bonus for SMP.

LearjetMinako


I haven't gotten one that I seen.  All I have been getting lately are the 8004 & 8049.

Is there a program that can track what WU and their PPD value you have done?  I know the GPUTracker did, but the FAH v7 doesn't.

 
LJ - to track what units you have done, along with TPF and PPD, use HFM.net. 
That program  tracks results and enables you to easily view the workunits you have done.
 
 
Tex, There are workunits that only come with the V7 client (from what I have seen), mainly A4 core.
If I run Tracker, I only pretty much only get 690x of the A3 variety.
 
even with the 10% bonus, that that P7647 is still significantly lower ppd than other A4 core projects like 8049, 8012 (which is pretty much all I have been getting on the V7 client) which normally come out at 30k on my rig.
I think the longer a workunit runs for, you should get a degree of additional point compensation.  This unit took over a day to complete, yet produced 20% less points than completing multiple smaller units.
 
 

I've been getting these WUs non stop for a while now :/

12:47tpf = 21.33 hours approx.
I'm thinking with the 10% it's more like 24-25k ppd. A little low? Yes. Outrageous? No

Just picked up WU 7647 other brother, the WU 7646.  And it had to be picked up by my Mercury rig, not quite suited to handle heavy long term WU.  Currently the i5 3570k is going threw it at normal speed of 10k PPD, 21min 31sec TPF.  The project should be done within about 36 hours. 

nathan_P

These 7647's remind me of the dreaded 6701's from a couple of years back - lots of work for little reward

Don't say that number very much...I'm not even gonna utter it, 'cause it might come back!! 

yodap


12:47tpf = 21.33 hours approx.
I'm thinking with the 10% it's more like 24-25k ppd. A little low? Yes. Outrageous? No

 
I agree with you. The problem is the incredible ppd variance, my ppd varies between 11k and 27k (A4 core) --> Benchmark topic

p7645 GRO-A4
2250pts
10.4days preferred
17.3 days deadline
 
tpf 9:40mins ...currently on my rig 
est credit 22748
est PPD 34k
 
WOW!!!...
a real bummer going from Bigadv doing 130K+ to doing 35k
Someone has got to get the message across on fixing the disparity between SMP and Bigadv Points
 
it wouldn't be so bad ...but before this I was getting 50-70k on SMP
so who dropped the low-point units on us?  (forgive me:  But I referred to them far worse earlier)
 
I read the Beta team thread on them....were not the only ones seeing an issue here
 
 
 

Xavier Zepherious

Someone has got to get the message across on fixing the disparity between SMP and Bigadv Points

Seems that PG wants us to funnel those "messages" via our DAB rep.  Do we still have one that is active, or maybe a better question, is the DAB still active?

I'm taking my daily driver off of v7 until these are gone. Not worth it for me and i'm chewing less WU than I would under v6.
 
Leaving my dedicated 5640 rig on them for now but i'm still down about 8k PPD

I believe rjbelans is our rep.
 
A thread on Wu with points issues for rj to bring up would be good
even someone on the beta team would be able to take it up
 

Xavier Zepherious

I believe rjbelans is our rep.
A thread on Wu with points issues for rj to bring up would be good
even someone on the beta team would be able to take it up

Maybe post something onto the tail-end of this thread since it is for DAB-related things and RJ was in there commenting just this past June.
 
http://www.evga.com/forum...48&mpage=1#1623130
 

Well, I was finally bestowed a 7647 last night on my 2600K @4Ghz.  My eyes about popped out when I saw how long it was going to take to complete.  I am also Folding on a GTX580 on that same box, so that has some affect, but this thing feels like a mini-Bigadv without the points/performance. This even puts the dreaded 6701 (of old) to shame.
 
My stats with this WU currently at 67%:
TPF = 09:07
Points (estimated) = 18,560
Time to complete (estimated) = 24.3hrs
PPD (estimated) = 18,285 (on a rig that typically does between 21-24K SMP with the GTX580 also Folding)
 
I Fold what they send me, but did have to report back that this thing finally found me...

My exact thoughts, been thinking this for quite some time.
I never could understand why they don't give a bonus for GPU folding like CPU folding?
Heck like you said some of these GPU WU's are like folding a -Bigadv on a GPU.
 

QRB for GPU's...it's on their list of things to do
 
I sooner have SMP point system and benchmark system fixed.
right now it's a little unbalanced with too much favoritism for bigadv
 
They could also bring in SMP 12 for i7 (bring back something for those that lost bigadv).
 
my conjecture - stanford has lost sight of their user base and are not listening to the DAB.
yes they have science to do...but not at the cost of making those doing their work mad and leaving
 
they need to remember the users and our work and energy in this and to keep the whole point system fair
 

Here you go there is currently an active discussion about these over at the FF . I currently get about 50% of what the best smp WU's give with these and they are about 30% below the avg smp for me. Any how this is not the first time it has been pointed out and probably will not be the last.
Hmm I can not post links the name on the thread over at the FF is Points drop off

http://foldingforum.org/viewtopic.php?f=67&t=22360&p=223482&hilit=Points+drop+off#p223482
 
you need 50 posts at EVGA to post links

Guess I do not have enough then.
Anyway as you can see over at the FF I am once again wrong. But then again I always am wrong.

just look at my stats
http://folding.extremeoverclocking.com/user_summary.php?s=&u=541641
 
3930k produces about 50k PPD(I was doing 55k) before all of this
my other rig is a constant 20k (gtx570+ q6600) gives approx 18-22k
 
it's now taking both rigs to maintain around 50K avg
that be a 30% drop
 
 

on a side note I was censored at FF for a post in a thread on the points issue (and point drop off)
 
instead of letting people have a say and let others comment and see if there is an issue - they censor
 

Yep, it was a matter of time before my rigs started to get murdered by these WU's.  Due to the long runtime per WU, I'm not letting Caboose fold anytime soon. 

Xavier Zepherious

on a side note I was censored at FF for a post in a thread on the points issue (and point drop off)
instead of letting people have a say and let others comment and see if there is an issue - they censor

I've noticed that too over there for a long time now.  The window of tolerance for most points-based discussion and a few other Folding related topics is very narrow or non-existent.  They lock threads very easily and then spout things like "we only discuss certain Folding topics" here.  The arrogance of answers by some regular people (that I'm sure you know over there) is also unnecessary and only dampens healthy discussion.  It appears they have certain people that think they are the self-appointed voice for PG, but they don't ultimately do PG any favors.  But PG allows them to run loose over there and the only time you know they are out of line is when Dr. Pande steps-in an offers his opinion that can be quite different.  I occasionally check that site just to stay informed, but it is not an forum atmosphere that encourages open-discussion IMO.

Little bump... Are you guys still getting these non-stop? I must have gotten 6 of them this week (considering they take about 16 hours...)

Between these and the 10125's, my SMP8 average is down to around 14k ppd from 20-22k. That's at 4 ghz!