Low ppd any idea why?

[xerxes242]

i just reinstalled windows xp on my computer and my ppd went from 5000 to 1500. i am using one core for the cpu client and the other for the gpu client. i have a e7200, 9600gt and i switched from a g31 gigabyte motherboard to a evga 750i sli motherboard when i reinstalled. i know its not an affinity problem because my cpu is always at 100%. also i found this in my log file.

[21:11:17] - Previous termination of core was improper.
[21:11:17] - Files status OK
[21:11:18] - Expanded 68499 -> 357580 (decompressed 522.0 percent)
[21:11:18] Called DecompressByteArray: compressed_data_size=68499 data_size=357580, decompressed_data_size=357580 diff=0
[21:11:18] - Digital signature verified
[21:11:18]
[21:11:18] Project: 5761 (Run 13, Clone 9, Gen 30)

could this have something to do with it?

thanks

[bruce]

could this have something to do with it?

Could what have something to do with it?

The message "Previous termination of core was improper" indicates that FAH crashed, but you didn't post that part of the log, so it's hard to say what happened.

[xerxes242]

this is what was above that part of the log.

--- Opening Log file [December 7 21:11:07 UTC]


# Windows GPU Console Edition #################################################
###############################################################################

Folding@Home Client Version 6.20

http://folding.stanford.edu

###############################################################################
###############################################################################

Launch directory: C:\Documents and Settings\Administrator\Application Data\Folding@home-gpu


[21:11:07] - Ask before connecting: No
[21:11:07] - User name: xerxes242 (Team 0)
[21:11:07] - User ID: 273B993D25AD2D8B
[21:11:07] - Machine ID: 2
[21:11:07]
[21:11:07] Loaded queue successfully.
[21:11:07] Initialization complete
[21:11:07]
[21:11:07] + Processing work unit
[21:11:07] Core required: FahCore_11.exe
[21:11:07] Core found.
[21:11:07] Working on queue slot 02 [December 7 21:11:07 UTC]
[21:11:07] + Working ...
[21:11:08]
[21:11:08] *------------------------------*
[21:11:08] Folding@Home GPU Core - Beta
[21:11:08] Version 1.19 (Mon Nov 3 09:34:13 PST 2008)
[21:11:08]
[21:11:08] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[21:11:08] Build host: amoeba
[21:11:08] Board Type: Nvidia
[21:11:08] Core :
[21:11:08] Preparing to commence simulation
[21:11:08] - Ensuring status. Please wait.
[21:11:15] Opening C:\Documents and Settings\Administrator\Application Data\Folding@home-gpu\MyFolding.html...
[21:11:17] - Looking at optimizations...
[21:11:17] - Working with standard loops on this execution.

[xerxes242]

i didn't notice the end of your post before. i meant could the error in the log file be the problem or could it be something to do with the new motherboard or reinstalling windows.

[ferrari12508]

my guess is you are getting a bunch of 511 point WUs for your GPU which is just causing it to get so few points. My 9800GT thats clocked as high as it can go while being stable will drop to about 3500PPD from 6000PPD

[Tigerbiten]

If your only running at ~1,500 PpD not ~5,000 PpD then it sounds like your card is running at 2D speeds.
If it is then your vid card drives may have a slight glitch.
Try uninstalling all the vid card drivers, run a drive cleaner, then reinstalling your drivers.

Luck .........

[P5-133XL]

You also should be checking to see if you are getting lots of errors on that card. I've seen cards that have been working just fine suddenly start getting continious EUE's (for some unknown reason) or get a WU that fails; does not report back that it has failed; and the server keeps giving the same bad WU over and over till the WU expires. Both of these senarios will result in your PPD dropping precipitiously for a while.

[toTOW]

Make sure your GPU client is set at higher priority than the CPU or SMP one (GPU : low or slightlyhigher and CPU/SMP : idle/lowest as possible).

[bruce]

I'm still not sure what problem you're reporting.

If you're concerned about the messages at [21:11:17], some of which are in each section of the log that you posted, they indicate that the previous time the client was running, it crashed. Usually that's immediately above the parts that you posted, but apparently it just got moved to FAHlog-Prev.txt so there may be information there about what happened. (Just look near the bottom of that file unless it contains a lot of errors.)

[xerxes242]

what i was trying to ask was could the improper termination message have something to do with the low points. i have not had any crashes so i am not sure what that is about. the clients are always still running when i check on them everyday. where would i find FAHlog-Prev.txt i did not see it and i am not able to search for files on this computer.

as for the other replies my priority is slightly higher on the gpu and lower on the cpu. i have been getting a lot of 511 and 384 point work units instead of the usual 480 point ones i get most of the time. are these just alot more time consuming per point?

[bruce]

Improper termination can cause reduced PPD but only if you're running the uniprocessor client. Both the GPU and the SMP clients always run their most optimized code.

[xerxes242]

i am running the uniproccessor client on one core and the gpu uses the other core. does improper termination refer to a process that the software is doing or is it termination of the whole program?

[bruce]

If you shut down the client "properly" such as with a CTRL-C for the Console client or a Quit in the console/GUI cliens, it should notify the FahCor_xx process(es) and they'll close all their files and shut down. If you have a BSOD or you kill the process (or, for that matter, if you kill the power), the system will not have time to close all the files properly.

[xerxes242]

i always just hit quit in the menu and haven't really had any bsod's and don't randomly unplug my computer so i'm not sure what that is from but thanks for the info.

[anko1]

For the CPU client, if it ends "improperly," when it restarts, it uses standard loops and production will drop. You can fix this by closing the CPU client properly (control C for console or Quit for systray) and then restarting. You'll want to see language like this after the WU starts:

[21:09:21] Project: 2484 (Run 149, Clone 24, Gen 7)
[21:09:21]
[21:09:22] Assembly optimizations on if available.
[21:09:22] Entering M.D.