F@H, Large Work Units for Server-class Available Now

I've been running 8106 for a little while now. And yes, PPD appears right in line with normal SMP. I'm on v7 with bigbeta and max-packet-size big. 2P 2670 V1.
 
Thanks for the info. If it turns out the same for my systems, which is logical given the 3.9K factor, I just won't be able to justify 675 Watts/200K PPD.
 
Yha, i'm on v6 -bigbeta and getting 58k ppd.

Looks like I might be dropping.
 
How could they possibly have released the server class units into the wild, letting all of us have a taste of bigadv class PPD again, then Nerf the points a week later?

Lord have mercy. More ineptitude in case we haven't had enough of that over the years. Its amazing I'm still folding at all
 
texinga said:
@Scotty...:D...good to see that you haven't lost your touch. A fresh round of "Bigadvism" wakes the dead. :eek:
Click to expand...

When your expectations for me are so low, it is quite easy for me to exceed your expectations :cool:
 
Thanks for the info. If it turns out the same for my systems, which is logical given the 3.9K factor, I just won't be able to justify 675 Watts/200K PPD.
Click to expand...
I didn't want to 'dump' on Pande Group, so I won't; I'll just try to be matter-of-fact about it. Pande Group has expressed the relative scientific value of the new "many-core" project work units by the points and resulting PPD. Given that - yes, I've run one of 3.9 K factor WUs now - my machines' output has one-third the scientific value as before. At 675 Watts per computer for 200K PPD each, it no longer makes charitable economic sense to me. (My power bill is NOT SMALL.) The machines are now turned off. I still want to give. I'm looking for a new charity or looking to sell this gear and get new tools. To those of you still multi-core folding, I salute you. To those of you who've chosen new pursuits, you have my respect as well.
 
I'm going to stick with regular SMP, points are the same and you actually far get more work units done, although 1 rig is currently off. Given the system requirements to run these large SMP a points premium over normal SMP is justified. I cannot see many of these WU getting folded for the current reward. Not when I can slap a pair of 980's in one machine and get over 700k PPD for less power than I will be burning on 2 machines
 
Let see what your reddit posting will bring ... :p

Else agree: Xeons really make a great GPU platform with a hell of PCIe lanes. Put some Maxwell's in (ok, still gambling) with reiser cards to give more air to breath or a bunch of 780(TI) if the energy bill can take it.

Still hope that PG can tweak the current concept a bit to not loose more donors with high class Xeons.
 
Yes it was to the point wasn't it:D

As for PCIe slots my shut down rig has space for 4 double slot gpu's or 7 if watercooled but with a 800w power limit at the moment i'll only be fitting one or 2 at most.

PG have already lost far too many donors with high end xeons/opterons, this will be the final nail in the coffin for most of the remainder
 
texinga said:
@Scotty...:D...good to see that you haven't lost your touch. A fresh round of "Bigadvism" wakes the dead. :eek:
Click to expand...

sc0tty8 said:
When your expectations for me are so low, it is quite easy for me to exceed your expectations :cool:
Click to expand...

Naw, I can go much lower than that...or would that be elevating the expectation? :D

Back to the subject at hand. The points (and associated costs to earn same) remains an unresolved problem (for PG) when Folding Enthusiasts get involved. I always felt that PG could hear the concerns of enthusiasts, but never really felt comfortable knowing how to manage it vs all the other (less invested) donors in the pond.

I was initially encouraged (and wary) at the thought that PG was rethinking their approach to recognize the SMP enthusiasts again. But, that appears to just be an error in points assignment and the same old direction as before.
 
Thanks for sharing your PPD experience, I was almost tempted to give it a try, but I guess given current PPD my xeons will keep crunching
 
Looks like you guys have more luck the I.

Tried throwing the v7 client on my 2P machine, and it's been putting out just under 8k ppd in the last 24 hours :)

I guess the 2P will stay dedicated to other assignments.
 
runs2far said:
Looks like you guys have more luck the I.

Tried throwing the v7 client on my 2P machine, and it's been putting out just under 8k ppd in the last 24 hours :)

I guess the 2P will stay dedicated to other assignments.
Click to expand...

Something is wrong somewhere, however the v6 client will work just fine
 
really disappointed with these new WU's

Going from an avg of 9 million PPD to 3.8 million PPD (even though I've added more machines to folding)

Quite a gut punch to go from overtaking Anonymous to falling behind :(
 
I'm seen a few

8106 - Avg. Time / Frame : 00:07:06 - 219,601 PPD
8107 - Avg. Time / Frame : 00:08:53 - 220,943 PPD

Down from:
8101 - Avg. Time / Frame : 00:13:08 - 364,964 PPD
8102 - Avg. Time / Frame : 00:09:18 - 612,474 PPD
8103 - Avg. Time / Frame : 00:09:14 - 619,119 PPD
8104 - Avg. Time / Frame : 00:07:08 - 602,547 PPD
8105 - Avg. Time / Frame : 00:09:20 - 594,954 PPD

All on basically 580 Watts
 
I'm going to fire up the 3rig tomorrow and test out one of these units, see what return I get on the 2p's.
 
You must log in or register to reply here.
Back
Top