PPD on avg going down?

[Shinare]

Hey all, I've noticed a trend on my folding farm thats alarming. I used to get about 2k-2.5k PPD out of each node but I'm not getting anywhere near that anymore, its been like that last couple weeks. Anyone else notice this or is it just me?

[just brew it!]

Yes, I have noticed as well. At least in my case, it appears to be due in large part to the fact that I am getting a lot of SMP WUs that use the a1 core instead of the a2 core.

[Patonb]

Yha, I've been nailed by this too, 2 and 3 day units, where there use to be mainly 15hr ones.

[notfred]

Yup, seeing the same as well. I'm even getting given 4 processor a1 units on my core i7 920 which usually does very nicely on 8 processor a2s which drops me by about 6k PPD.

[Shinare]

Yah, its just killin' me..

My wife made my buy furniture instead of the stack of 4850's I had in my wishlist. *cry*

[Bruce]

I'm getting lots of "Attempt #xx to get work failed, and no other work to do", and have about a quarter of my available cores twiddling their thumbs

Bruce

[silent ninjah]

I get that for about 10 attempts a try on my single core clients. GPU never has an issue.

It's also unable to upload a bunch of WU's from the same period, error 503, but afaik that just means the server isn't accept/giving out WU's anymore, so I think I've lost those points.

[Krogoth]

It seems that some F@H WU servers are having problems in the past month.

I have been encounter connection issues with only certain Project groups, while others are working fine.

[Momento Mori]

My SMP client has had diminishing returns as well recently, but the GPU clients are keeping my PPD high.

[farmpuma]

Yep, the single core client has been getting the redheaded bastard stepchild treatment from the servers for several weeks now. And lately it's been worse. The majority of the time the client connects to the server, but fails to be given a work unit because the server is to busy or having issues.

Also, I have three (on three different machines) DGromacs from 128.59.74.4 which has been in fail/reject mode for over a week or more. Plus I had two DGromacs (on two different machines) from 128.143.48.226 fail to upload because the server didn't like them, but gave no reason why they were defective. Both work units ran with no errors or other indications of a problem. And both of the machines had finished and uploaded several of the same work units with no issues. I suspect, "garbage in, garbage out."

The GPU server has been sending out the 511 point sloow on Nvidia work units the majority of the time for a while and has lately mixed in some also slowish on Nvidia 420 point work units.

On the upside, and I hesitate to mention it, the SMP server has been treating both of my 5.91 windows clients quite well lately.

[Flying Fox]

Yep, the single core client has been getting the redheaded bastard stepchild treatment from the servers for several weeks now. And lately it's been worse. The majority of the time the client connects to the server, but fails to be given a work unit because the server is to busy or having issues.

The fatter Gromacs C WUs seem to have dried up and I have been getting some slow but small ones. I also have a WU stuck for more than 2 weeks to the point I just said "screw it" and trashed it.