FAH GPU Tracker V2 is a Folding@Home Client tracking and control program that I created to make folding easier.

Features:
- Controls up to 9 clients (CPU or SMP, up to 8 GPUs)
- Tracks detailed stats for every client and every name/team combo
- Tracks WU failures by project number for each client
- Saves the logs of failed WUs
- Can automatically reconfigure user information from an XML file
- Gaming Pause feature to pause clients while gaming
- Automatically updating projects list (pulled from psummary)
- Hides all client windows to prevent taskbar clutter
- Can be minimized to system tray
- Automatic update support
- Heat Control setting to throttle GPUs based on the current WU
- Allows switching between the GPU2 and GPU3 clients (per-GPU)
- Can export an XML file with the current status of the Tracker
- Can upload the XML status file to an FTP server
- Downloads the FAH clients from Stanford on first run
- PPD monitoring of all clients including bonus points for SMP A3

Download v3.54 - Final:
FAH GPU Tracker V2
 
Setup Guide:
How to start folding - the easy way

FAH GPU Tracker V2 website:
FAH GPU Tracker V2 Website

Support Forum:
FAH GPU Tracker V2 Support Forum

Installation:
1. Extract FAH GPU Tracker V2.zip
2. Run FAH GPU Tracker V2.exe (If you have UAC enabled run it as administrator)
3. Go to Setup -> Configure
4. Change the options as you see fit
5. Hit apply and close the config page
6. You are ready to start folding!
7. Hit the start button(s) on the desired client(s) to begin folding

Excellent News!  I'm very happy you did this!  Blue Ribbon requested....
 
One of my favorite things about this utility is the ability to switch to -oneunit mode without having to manually stop and restart the clients.
 
Edit/Requests:
1)    On the version you did last year, it let me chose -oneunit mode per client, but it appears that in this version the -one unit flag is global, could you make the -oneunit flag per client?
 
2)    Any chance you could modify the Tracker window to only display the clients we've enabled?  For example, if one only has two GPU clients running, it doesn't display the other 6 GPU clients and 2 CPU clients?

I'd also like to recommend this utility for another biq reason - It installs all the folding clients you need, so you don't have to download and set them all up seperately, and puts each in it's own nice sub-folder to make it easy to keep track of everything.  I think this makes it excellent for newcomers to folding.

Davabled


Excellent News!  I'm very happy you did this!  Blue Ribbon requested....
 
One of my favorite things about this utiity is the ability to switch to -oneunit mode without having to manually stop and restart the clients.
 
Edit/Requests:
1)    On the version you did last year, it let me chose -oneunit mode per client, but it appears that in this version the -one unit flag is global, could you make the -oneunit flag per client?
 
2)    Any chance you could modify the Tracker window to only display the clients we've enabled?  For example, if one only has two GPU clients running, it doesn't display the other 6 gpu clients and 2 cpu clients?

1) This is a simple change, I will release an update in a bit with this changed.

2) This is fairly difficult to do with the current way I have it setup in VB. I could hide the other clients, but it would just make them invisible, it would not move the other clients or resize the main window. The best I am able to do at the moment is set the status to "Disabled"

dl'd....but i am trying to set up the smp client now.

Great post! 

Is the EOCid the same for everyone?

Impressive work.  Looks to be much more noob friendly than everything else out there.  They should give ya a blue ribbon, a t-shirt and a bumper sticker.  LOL

Can't get the SMP client running.

clo007


Is the EOCid the same for everyone?

No, this is your ID for the Extreme Overclocking Stats site. This is an optional setting that just makes the "Stats -> View EOC Stats" option work. It can be found at the end of the URL for your EOC stats page.

For example, my EOC stats page is:
http://folding.extremeove...ry.php?s=&u=443273

The EOC ID is the last part of the URL, so my EOC ID is 443273.

clo007


Can't get the SMP client running.

Did you run the install.bat file in the SMP client folder? What OS are you using?

yes. win7 64bit. I launch and it takes me to the black dialog box. I am not sure what info I need to put in.

clo007


yes. win7 64bit. I launch and it takes me to the black dialog box. I am not sure what info I need to put in.

This is the same install.bat used for the standard SMP client. How to:

1.Basically when it comes up the first step is to enter the account name of an admin account on your system (if you are and admin just hit enter to accept the current account name)
2.Once you have done that, you will have to enter the password to the account you specified in the first step twice (it will not show anything being entered, this is normal).
3. If you completed the first 2 steps correctly you should see "If you see this twice, MPI is working" 2 times.

got it jedi thanks man!!!

This is awesome man. Big thanks for doing this. I now got my GTX275 folding, as well as my Q9550...and you made it so damn easy!!!

SMP client seems slow...especially 4 cores at 4.1ghz

clo007


SMP client seems slow...especially 4 cores at 4.1ghz

What are you using to determine PPD? FahMon doesn't calculate the bonus points for the core A3 WUs. I recommend HFM.net as it can correctly calculate bonus points. If you do not have a passkey you will not receive the bonus points. You can get a passkey here:
http://fah-web.stanford.edu/cgi-bin/getpasskey.py

In order to get bonus points you also have to have at least 10 A2 or A3 WUs completed, and have >=80% of all A2 and A3 WUs completed within the deadline.

I am just monitoring the progress from your program. GPU is at 44% complete while the SMP is only at 8%

http://img442.imageshack.us/img442/8958/screeniec.jpg

SMP work units are much larger than most GPU work units, so it may seem slower, but that is just because it is bigger (and worth more points).

Based off what i am reading, it actually worth less points, so I don't know what the deal is.

clo007


Based off what i am reading, it actually worth less points, so I don't know what the deal is.

The base credit of the A3 core SMP WUs is very low without the bonus points. The Tracker does not calculate bonus points at this time.

oh ok...well I am just glad i got my gfx and cpu folding...your program made it super easy and you should get multiple Br's for this...and eVGA should support it for all folders as well. I don't know how they could, but they should.

I did run into one problem running multiple GPU clients that I didn't have the old fashioned way.

The problem:  If one GPU client fails (resulting in a Display Driver Stopped Responding error), it crashes the GPU Tracker program as well.  I was a little too impatient and can't confirm if this meant the other client(s) controlled by GPU Tracker continued folding in the background or not.
 
Edit: I'm running Win 7 Ult. 64-bit RTM, a pair of gtx260's with driver version 186.18

Davabled


I did run into one problem running multiple GPU clients that I didn't have the old fashioned way.

The problem:  If one GPU client fails (resulting in a Display Driver Stopped Responding error), it crashes the GPU Tracker program as well.  I was a little too impatient and can't confirm if this meant the other client(s) controlled by GPU Tracker continued folding in the background or not.
 
Edit: I'm running Win 7 Ult. 64-bit RTM, a pair of gtx260's with driver version 186.18

Can you send me your LOG.txt file located in the main Tracker folder? I might have an idea on how to fix this. Also, if one of the clients continues to run after the Tracker closes, it will not longer be able to stop that client once it restarts until you stop it manually. You could also just wait for it to close on its own, which will happen once the WU is complete. This has never happened to me during normal operation, so it should only be an issue if the program crashes.

jedi95

Davabled


I did run into one problem running multiple GPU clients that I didn't have the old fashioned way.

The problem:  If one GPU client fails (resulting in a Display Driver Stopped Responding error), it crashes the GPU Tracker program as well.  I was a little too impatient and can't confirm if this meant the other client(s) controlled by GPU Tracker continued folding in the background or not.

Edit: I'm running Win 7 Ult. 64-bit RTM, a pair of gtx260's with driver version 186.18

Can you send me your LOG.txt file located in the main Tracker folder? I might have an idea on how to fix this. Also, if one of the clients continues to run after the Tracker closes, it will not longer be able to stop that client once it restarts until you stop it manually. You could also just wait for it to close on its own, which will happen once the WU is complete. This has never happened to me during normal operation, so it should only be an issue if the program crashes.

Here's the contents of the log file:
 

      [10:27:19] --------------------------------------------------
     [10:27:19] FAH GPU Tracker v2.0 Startup
     [10:27:19] --------------------------------------------------
     [10:29:23] Applied new config settings
     [10:29:47] Starting GPU 1
     [10:29:47] Started new WU on GPU1
     [10:59:31] Starting GPU 0
     [10:59:31] Started new WU on GPU0
     [11:01:33] GPU 0 failed to complete a project 10105 WU (EARLY_UNIT_END)     

 
 
After that I went back to running the clients manually.

So it crashed after that entry correct? If so I have a good idea of where to look for the cause.

jedi95


So it crashed after that entry correct? If so I have a good idea of where to look for the cause.

Correct.  After that message GPU Tracker crashed and Windows asked if I wanted to submit a crash report.

Davabled

jedi95


So it crashed after that entry correct? If so I have a good idea of where to look for the cause.

Correct.  After that message GPU Tracker crashed and Windows asked if I wanted to submit a crash report.

I have released a possible fix for this problem with v2.2. The main download link has been updated for anyone downloading the program for the first time. If you already have it you can update easily be going to "Setup -> Check for updates"

I like the fact that upon starting version 2.0, it let me know 2.2 was available and asked me to update... very nice.

However, I get the following error when trying to update: 
 

      [19:49:44] --------------------------------------------------
     [19:49:44] FAH GPU Tracker v2.0 Startup
     [19:49:44] --------------------------------------------------
     [19:49:47] Attempting to update FAH GPU Tracker V2
     [19:49:48] Update failed - downloading new version failed
     [19:50:01] Attempting to update FAH GPU Tracker V2
     [19:50:02] Update failed - downloading new version failed     

Davabled


I like the fact that upon starting version 2.0, it let me know 2.2 was available and asked me to update... very nice.

However, I get the following error when trying to update: 
 

      [19:49:44] --------------------------------------------------  
 [19:49:44] FAH GPU Tracker v2.0 Startup  
 [19:49:44] --------------------------------------------------  
 [19:49:47] Attempting to update FAH GPU Tracker V2  
 [19:49:48] Update failed - downloading new version failed  
 [19:50:01] Attempting to update FAH GPU Tracker V2  
 [19:50:02] Update failed - downloading new version failed     

I just had a facepalm moment when I looked at the code and found the problem. I had it trying to download the file from my local server instead of the proper site. For anyone who has version 2.2 or earlier you will need to manually download this file and replace FAH GPU Tracker V2.exe with the one in this archive:
<link removed, please download the latest full version>

Thanks for the quick updates, I'll give 2.21 a try shortly.

I just noticed a feature missing from the old version, the option to specify the SET variable for GPU clients based on Work Unit (for example, setting 511's to 50%, 353's to 100%, 478's to 90%, etc...)

That option came in VERY handy since specific work units tended to be very hot, while others using the same core would run fine.  Something we see talked about a lot in this and other forums, I believe many of our folders might switch over just for that functionality alone.  Hopefully the switch to using psummary vs. a hard-coded WU index won't prevent this feature from coming back.