Breaking News: SMP, now with even more sucky PPD!

Skripka

[H]F Junkie
Joined
Feb 5, 2012
Messages
10,791
So PG has released more new SMP units...and well...I'll let this speak a thousand words:

*Note machine is 100% load with F@H getting all of it, smp only. Win7. And yes, CPU-Z is reading my system correctly. And yes, these TPFs are stable....and yes this is an i7-2600 getting 10k PPD.

PerbF0J.png
 
I'm running 100% on a stock i5 2500k, and I'm getting 8,000 - 10,000 PPD.

For me it seems like the larger the WU, the worse the PPD - the ones that take over a day to finish bring my PPD down to the 8,000 range. Shorter ones that finish in a couple hours bring my PPD up to almost 12,000. Maybe its the project, but it seems size matters.
 
fhcontrolppd.jpg


I see you're working on a 7087... I haven't had any of those yet to tell if it's just the project. Are you seeing low PPD on all WU's?
 
I have a pair of LGA771 systems. both running a pair of L5420's. Just checked my history and there are some of each project from October, some in December and one from the January 15th. They might have changed them and re-released.?

Project ID: 7646
Core: GRO-A4
Credit: 2250
Frames: 100


Name: FAH001
Path:
Number of Frames Observed: 200

Min. Time / Frame : 00:15:11 - 17,287.4 PPD
Avg. Time / Frame : 00:15:17 - 17,118.0 PPD


Name: FAH002
Path:
Number of Frames Observed: 300

Min. Time / Frame : 00:14:55 - 17,753.0 PPD
Avg. Time / Frame : 00:15:04 - 17,488.5 PPD



Project ID: 7087
Core: GRO-A4
Credit: 1157.48
Frames: 100

Name: FAH002
Path:
Number of Frames Observed: 200

Min. Time / Frame : 00:07:49 - 14,954.8 PPD
Avg. Time / Frame : 00:07:50 - 14,907.1 PPD
 
Well FWIW my 1090Tx6 gets 11K ppd on P7646, seems like a chewy unit. No P7087's so far ...
 
That's not good news at all. Hasn't been a very good trend the last few months :/ At least the WUs lately haven't been producing as much heat as the ones from a month or two ago. May I ask what you guys are tracking your WU history with?
 
HFM.net here. It keeps your entire WU history in a local database. Works with v6 or v7 clients and dead easy to set up. I have mine configured to export to a web page so I can check stats on the fly from my smartphone.
 
Damn you double posting!
 
Just got a 7007 on a laptop with an i3, an it's saying 3048 ppd.
I've got a 10450 on a desktop i3 and it's listed as 3533 ppd.
I'm not sure how to feel about this. >.>
 
This will certainly continue to encourage us to use GPUs rather than cpus, which is going to suck in the summer.
 
I dunno, its been a bad couple weeks for GPUs too for me. I've had a plethora of driver crashes and failed WUs. Been playing the driver musical chairs game.
 
I dunno, its been a bad couple weeks for GPUs too for me. I've had a plethora of driver crashes and failed WUs. Been playing the driver musical chairs game.

Nvidia? (from your sig) I've been using 306.97 on 10x GPUs (460/560Ti/570) with success. Got performance mode enabled? Try turning off the screen saver? What issues are you having in detail?
 
I think they are just reducing the points over time to compensate for faster CPUs
 
I think they are just reducing the points over time to compensate for faster CPUs

Except in the desktop arena there haven't been meaningfully faster procs released since Sandybridge was released at the very beginning of 2011....back then an i7-2600K could net you 80,000PPD or so?
 
Except in the desktop arena there haven't been meaningfully faster procs released since Sandybridge was released at the very beginning of 2011....back then an i7-2600K could net you 80,000PPD or so?

Only through neferious means that, as a team, we never condoned.
 
Only through neferious means that, as a team, we never condoned.

True...but the smp PPD drop is purely arbitrary on Stanford's part. Especially given the gift of ludicrously high GPU QRB that was tested...and is in all liklihood going to be implemented.
 
True...but the smp PPD drop is purely arbitrary on Stanford's part. Especially given the gift of ludicrously high GPU QRB that was tested...and is in all liklihood going to be implemented.

Oh I agree but i can't understand why it wasn't picked up in the beta test, I've been following new WU over at FF and more than one SMP project has had its points changed because they were too low. Personally i think that A4 WU have always given less points than an A3 unit
 
My i7-950 PPD @ 4ghz has been abysmal lately. 8k is hard to swallow for the amount of power it burns... and the fact that it used to be bigadv capable :(
 
I think Stanford is basically saying they only want heavy duty mult-cpu machines or GPUs crunching.
 
my 980x at 4.2 that usually brings in 30k+ is sitting at 6500 right now, not cool
 
I simply don't understand why they make these "older" systems have less PPD as newer systems come out. Shouldn't they just increase the PPD of the newer systems. 2 years ago that system was worth running, but today it might not be. I think PG is shooting themselves in the foot here. If they simply gave new WUs to newer systems with faster processors, but left the PPD for the old systems alone, more older systems would stay online.
 
Wow, I thought my computer was borked due to the smp getting low ppd. Did a reboot just to make sure.
 

It's funny you say they want more GPU foldings. I turned my 8800GTX on the otherday to see what would happen, and it netted me in the low 9K PPD.


I was shocked. I just hate that fan spinning on my video card so I let it finish a couple WUs and turned it off.
 
It's funny you say they want more GPU foldings. I turned my 8800GTX on the otherday to see what would happen, and it netted me in the low 9K PPD.


I was shocked. I just hate that fan spinning on my video card so I let it finish a couple WUs and turned it off.

Thats a lot better than you used to be able to get, I never got above 5k on mine whilst it was folding
 
Nvidia? (from your sig) I've been using 306.97 on 10x GPUs (460/560Ti/570) with success. Got performance mode enabled? Try turning off the screen saver? What issues are you having in detail?

OP: Sorry if I'm jacking your post, if you want me to take this elsewhere, let me know and I apologize in advance.

One of my rigs is a Q8400 with a 460 and a 550ti. A while back the 460 started crashing the driver and that was remedied by incrementing voltage up and core clocks down, eventually that stopped helping and it would just fail everything. I assumed the card was bad, but as a precaution I tried different driver versions, eventually finding one that was stable. It allowed me to reset the card to stock. That lasted a month or two, and the problems progressively came back, I compensated with volt and clock tweaks. About a week ago it started failing all WUs again, except this time it was also taking out the driver and crashing fahcore, thus stopping the 550ti. I worked my way up to the 310.xx whql driver, and all is stable again as of two nights ago. o_O
Now everything is running but my other rig with 480s is making 25K ppd per card on 7624s and 7625s... yuck.
 
Thats a lot better than you used to be able to get, I never got above 5k on mine whilst it was folding

I used to get 5k on my 8800gt ... on linux :D

It seems that GPU ppd is just as crazy as SMP ...
 
Anyone get a 7083 WU?

I've been overclocking my 2500k which upped my avg PPD to ~14-16000. This WU is getting 3500 :eek:
 
Open Task Manager. Click on the Processes and then click on the CPU field. FahCore_a3 (or _a4) how many percent shows? Should be about 99-97%. If it is less, see which other process(es) uses resources from the CPU.

task_manager_965e3_30370003.png
 
Yeah, it loads up to 98-99% on average. Anyway, I chugged through that WU and onto the next. PPD back up to normal.
 
All about the Nvidia GPU now. Gotta feel bad for the guys with the big cpu systems, but not for me :p
 
Back
Top