Guide to running multiple GPU work units

[ThE_MaD_ShOt]

Originally Posted by Norton

I'm peaking @87% load and 60C max temps so I can definitely fit in a 4th wu. The peak load% only lasts for 5-6 seconds and then it drops into the 70's due to the stagger on the wu's.

I think I can squeeze in a 5th wu without causing errors but I don't think I'll push it THAT hard...

Let it run for a bit. If the wu's fall on top of each other the load percentage can raise. But if it remains at 87% you just might be able to squeeze a fourth one in.



Here is 2 on a 7850. I probably have room for atleast 1 more but I'll just wait. Also both completed and went to upload state in 2mins 50 secs.

[brandonwh64]

Ok now since its ran for a few minutes my WU are estimating 5-7 minutes a piece

[ThE_MaD_ShOt]

Originally Posted by brandonwh64

Ok now since its ran for a few minutes my WU are estimating 5-7 minutes a piece

http://img.techpowerup.org/121021/Capture032.jpg

Cool just keep a close eye for awhile. and for the next day or so and make sure you don't see any errors. I have seen some people start erroring out after a day or so of running. But you do have so sweet cards so you will probably be okay.

You my friend are going to be back into pie mode.

My next project for next weekend is to see how many at a time a 2gb 5970 can do. I just need to keep a careful eye on that card for heat.

[brandonwh64]

Originally Posted by ThE_MaD_ShOt

Cool just keep a close eye for awhile. and for the next day or so and make sure you don't see any errors. I have seen some people start erroring out after a day or so of running. But you do have so sweet cards so you will probably be okay.

You my friend are going to be back into pie mode.

My next project for next weekend is to see how many at a time a 2gb 5970 can do. I just need to keep a careful eye on that card for heat.

Im going to stop it for the night but tomorrow morning I will start it back up before I goto work.

[BUCK NASTY]

Works for Nvidia! Got 2 running on a GTX460. Times went from 6:00 for 1 WU to 8:00 for 2 WU's. Thanks KieX!!!!!

Edit: Down to 7:30 per WU.

[KieX]

Originally Posted by BUCK NASTY

Works for Nvidia! Got 2 running on a GTX460. Times went from 6:00 for 1 WU to 8:00 for 2 WU's. Thanks KieX!!!!!

http://img.techpowerup.org/121021/Capture033.jpg

That's good to know, thanks for testing it. Think I'll try it out with the 670 and 560 tomorrow.

[Norton]

All of my gpu wu's just vanished

Deleted the app_info file and restarted the BOINC mgr... no idea what just happened???

[KieX]

Originally Posted by Norton

All of my gpu wu's just vanished

Deleted the app_info file and restarted the BOINC mgr... no idea what just happened???

This is the sucky thing about using this file. Tried as best to warn in the OP.. but changes to the file can result in loss of the previous WU's

EDIT: Modified the OP to have another Disclaimer at the start to be clearer on that point. Guess we learn on the trot with this.

[ThE_MaD_ShOt]

Okay guys apparently just shutting down Boinc, dropping the file into the folder and restarting Boinc is a no no. Norton and I both have had serious problems doing this. It will run the multiple Wu's but it would empty the buffer and would get new work no matter what we tried. If you a having these issues let me know and I will post up the fix for it.

For those just seeing this post and want to try the script, the proper way to load the app_info file is as follows:

The procedure to install an app_info.xml is to first hit the 'No new tasks' under 'Project' in the Boinc manager then you wait until all your workunits have finished and have been returned and after that you shut down BOINC and then you copy your app_info.xml file in to the BOINC\Data\projects\www.worldcommunitygrid.org map and restart BOINC. Then you hit 'Allow new tasks' in the BOINC manager and click 'Update' under 'Projects' in the BOINC manager, that's it.

If you recieve this error under messages:

10/21/2012 10:19:43 PM World Community Grid [error] App version returned from anonymous platform project; ignoring

Simply ignore it.

If you receive this error:

10/21/2012 10:19:43 PM World Community Grid [error] No application found for task: windows_intelx86 640 ; discarding

It indicates that you have not spell the name of the application in the app_info.xml correct or the application is missing in the BOINC\Data\projects\www.worldcommunitygrid.org map, if that's the case you can hit 'Reset project' under 'Projects' in the BOINC manager.


Hope this helps and like I said if you are having problems as I describe in the beginning let me know and I will tell you how to fix your boinc install.

[[Ion]]

Originally Posted by BUCK NASTY

Works for Nvidia! Got 2 running on a GTX460. Times went from 6:00 for 1 WU to 8:00 for 2 WU's. Thanks KieX!!!!!

Edit: Down to 7:30 per WU.

http://img.techpowerup.org/121021/Capture033.jpg

Woot!
I've just set both of my GPU systems to start emptying their caches, and I'll try this tweak once they've run out of WUs. Should be ~2 days on the i7-860 and ~3 days on the i7-2700k

Originally Posted by Norton

All of my gpu wu's just vanished

Deleted the app_info file and restarted the BOINC mgr... no idea what just happened???

You know when I was having a terrible time with one of my rigs a week or so back...this was the root of that issue. It can really make BOINC behave strange

[Bun-Bun]

Originally Posted by KieX

Driver crahses:
If you are having driver crahses, the following registry modification might prevent that:

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Watchdog]
"DisableBugCheck"="1"

[HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Watchdog\Display]
"EaRecovery"="1"

Looks like you took this from me over at XS!

Just kidding.

Anyway I had an error in it that I corrected in my OP at XS and thats is "EaRecovery"="1" should be ="0"

[KieX]

Originally Posted by ThE_MaD_ShOt

Okay guys apparently just shutting down Boinc, dropping the file into the folder and restarting Boinc is a no no. Norton and I both have had serious problems doing this. It will run the multiple Wu's but it would empty the buffer and would get new work no matter what we tried. If you a having these issues let me know and I will post up the fix for it.

For those just seeing this post and want to try the script, the proper way to load the app_info file is as follows:

The procedure to install an app_info.xml is to first hit the 'No new tasks' under 'Project' in the Boinc manager then you wait until all your workunits have finished and have been returned and after that you shut down BOINC and then you copy your app_info.xml file in to the BOINC\Data\projects\www.worldcommunitygrid.org map and restart BOINC. Then you hit 'Allow new tasks' in the BOINC manager and click 'Update' under 'Projects' in the BOINC manager, that's it.

If you recieve this error under messages:

10/21/2012 10:19:43 PM World Community Grid [error] App version returned from anonymous platform project; ignoring

Simply ignore it.

If you receive this error:

10/21/2012 10:19:43 PM World Community Grid [error] No application found for task: windows_intelx86 640 ; discarding

It indicates that you have not spell the name of the application in the app_info.xml correct or the application is missing in the BOINC\Data\projects\www.worldcommunitygrid.org map, if that's the case you can hit 'Reset project' under 'Projects' in the BOINC manager.


Hope this helps and like I said if you are having problems as I describe in the beginning let me know and I will tell you how to fix your boinc install.

Will update OP accordingly.

Originally Posted by Bun-Bun

Looks like you took this from me over at XS!

Just kidding.

Anyway I had an error in it that I corrected in my OP at XS and thats is "EaRecovery"="1" should be ="0"

I put a link back to the original thread, not a member at XS so couldn't give you a thanks. Happy to put your name in though, this isn't school plagiarism.

Thanks for posting and correcting that error, I'll make the changes needed

[Bun-Bun]

Haha no worries. I just found it amusing I found your thread while researching further into the topic and it had the same error mine did.

Glad the information is being seen by more of the WCG community.

[KieX]

Originally Posted by Bun-Bun

Haha no worries. I just found it amusing I found your thread while researching further into the topic and it had the same error mine did.

Glad the information is being seen by more of the WCG community.

Thanks! Keep up the good work

[Bun-Bun]

In case anyone is curious this is where it came from

http://msdn.microsoft.com/en-us/libr...(v=vs.85).aspx

Could also be the solution to the problem many have with both Nvidia/ATI/AMD and the timeout error/BSODs they have been getting. I know it has plagued me in the past and I would have loved to known about this issue.

[BUCK NASTY]

Originally Posted by ThE_MaD_ShOt

10/21/2012 10:19:43 PM World Community Grid [error] No application found for task: windows_intelx86 640 ; discarding

I got this error and found the rig idling this morning. De-tached and re-attached and back to a single work unit. How can i fix this MaD_ShOt?

[ThE_MaD_ShOt]

Pmed' you Buck as I needs some more info fro you and instead of posting a buch of threads in here we will get it taken care of in pm. Then if I see you have the same problem as me and Norton I will post the fix in here.

[Jstn7477]

I tried the app_info.xml on my 7950 with the default 3 WUs setting and Catalyst 12.11 instantly crashes. I looked for the registry settings but they don't seem to be there (no Watchdog folder) on Windows 8. Help?

[KieX]

Originally Posted by Jstn7477

I tried the app_info.xml on my 7950 with the default 3 WUs setting and Catalyst 12.11 instantly crashes. I looked for the registry settings but they don't seem to be there (no Watchdog folder) on Windows 8. Help?

Which version of BOINC are you running? Just wondering if it's got something to do with this: http://www.worldcommunitygrid.org/fo...d_thread,33527

What error messages come up in the Event Log?

[Jstn7477]

Originally Posted by KieX

Which version of BOINC are you running? Just wondering if it's got something to do with this: http://www.worldcommunitygrid.org/fo...d_thread,33527

What error messages come up in the Event Log?

7.0.28 x64 from Berkeley. I tried the old WCG one initially but it wouldn't even install.

Events are 4101 (Warnings) saying "Display driver amdkmdap stopped responding and has successfully recovered."

[KieX]

Originally Posted by Jstn7477

7.0.28 x64 from Berkeley. I tried the old WCG one initially but it wouldn't even install.

Events are 4101 (Warnings) saying "Display driver amdkmdap stopped responding and has successfully recovered."

I'll browse around see if I can come up with anything. Have you tried rolling back to 12.8 Catalyst see if that resolves the issue?

[Norton]

Originally Posted by KieX

I'll browse around see if I can come up with anything. Have you tried rolling back to 12.8 Catalyst see if that resolves the issue?

I'm running on 12.6's and not having any issues with the 6870 or the 7870....

[mstenholm]

Originally Posted by ThE_MaD_ShOt

Pmed' you Buck as I needs some more info fro you and instead of posting a buch of threads in here we will get it taken care of in pm. Then if I see you have the same problem as me and Norton I will post the fix in here.

Add me to the list of people running out of GPU WUs.

Edit: I emptied the buffer, removed Boinc, installed it again and added the file and it looks OK.

[Mindweaver]

I put the file in and I'm crunching 2x GPU WU's but no CPU WU's? wtf? I reset the project.. but still no CPU work...

10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app sn2s platform windows_intelx86 ver 620 class ; discarding SN2S_AAG25601_0000168_0238_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app sn2s platform windows_intelx86 ver 620 class ; discarding SN2S_AAG25601_0000168_0271_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app sn2s platform windows_intelx86 ver 620 class ; discarding SN2S_AAG25601_0000168_0264_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app dsfl platform windows_intelx86 ver 625 class ; discarding DSFL_00020-47_0000036_0339_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app dsfl platform windows_intelx86 ver 625 class ; discarding DSFL_00020-47_0000036_0801_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app hfcc platform windows_intelx86 ver 640 class ; discarding HFCC_target-10_00782401_target-10_0001_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app hfcc platform windows_intelx86 ver 640 class ; discarding HFCC_target-10_00782626_target-10_0001_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app hfcc platform windows_intelx86 ver 640 class ; discarding HFCC_target-10_00785022_target-10_0000_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app hfcc platform windows_intelx86 ver 640 class ; discarding HFCC_target-10_00785299_target-10_0001_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app hpf2 platform windows_intelx86 ver 640 class ; discarding qi388_00116_18

Also, I'm using a different configuration I found over at XS, but the one listed in this OP didn't work for me either. I choose this other one because it has x64 apps in it.

   <app>
      <name>hcc1</name>
      <user_friendly_name>Help Conquer Cancer</user_friendly_name>
   </app>
   <file_info>
      <name>wcg_hcc1_img_6.56_windows_intelx86__nvidia_hcc1</name>
      <executable/> 
   </file_info>
   <file_info>
      <name>hcckernel.cl.6.56</name>
      <executable/>
   </file_info>
      <app_version>
      <app_name>hcc1</app_name>
      <version_num>656</version_num>
      <platform>windows_intelx86</platform>
      <plan_class>nvidia_hcc1</plan_class>
      <avg_ncpus>1.0</avg_ncpus>
      <max_ncpus>1.0</max_ncpus>
      <coproc>
         <type>CUDA</type>
         <count>0.5</count>
      </coproc>
      <file_ref>
          <file_name>wcg_hcc1_img_6.56_windows_intelx86__nvidia_hcc1</file_name>
          <main_program/>
      </file_ref>
      <file_ref>
         <file_name>hcckernel.cl.6.56</file_name>
         <open_name>hcckernel.cl</open_name>
      </file_ref>
   </app_version>
<app>
    <name>sn2s</name>
    <user_friendly_name>Say No to Schistosoma</user_friendly_name>
</app>
   <file_info>
      <name>wcgrid_sn2s_vina_6.20_windows_x86_64</name>
      <executable/>
   </file_info> 
   <file_info>
      <name>wcgrid_sn2s_vina_prod_x86_64.exe.6.20</name>
      <executable/>
   </file_info>
   <file_info>
      <name>wcgrid_sn2s_gfx_prod_x86_64.exe.6.20</name>
       <executable/>
   </file_info>
<app_version>
    <app_name>sn2s</app_name>
    <version_num>620</version_num>
    <platform>windows_x86_64</platform>
    <avg_ncpus>1.0</avg_ncpus>
    <max_ncpus>1.0</max_ncpus>
    <flops>3347548492.458962</flops>
    <api_version>7.1.0</api_version>
    <file_ref>
        <file_name>wcgrid_sn2s_vina_6.20_windows_x86_64</file_name>
        <main_program/>
    </file_ref>
    <file_ref>
         <file_name>wcgrid_sn2s_vina_prod_x86_64.exe.6.20</file_name>
          <open_name>AutoDockVina64.exe</open_name>
     </file_ref>
     <file_ref>
         <file_name>wcgrid_sn2s_gfx_prod_x86_64.exe.6.20</file_name>
          <open_name>graphics_app</open_name>
     </file_ref>
</app_version>
<app_version>
    <app_name>sn2s</app_name>
    <version_num>620</version_num>
    <platform>windows_intelx86</platform>
    <avg_ncpus>1.0</avg_ncpus>
    <max_ncpus>1.0</max_ncpus>
    <flops>3347548492.458962</flops>
    <api_version>7.1.0</api_version>
    <file_ref>
        <file_name>wcgrid_sn2s_vina_6.20_windows_x86_64</file_name>
        <main_program/>
    </file_ref>
    <file_ref>
         <file_name>wcgrid_sn2s_vina_prod_x86_64.exe.6.20</file_name>
          <open_name>AutoDockVina64.exe</open_name>
     </file_ref>
     <file_ref>
         <file_name>wcgrid_sn2s_gfx_prod_x86_64.exe.6.20</file_name>
          <open_name>graphics_app</open_name>
     </file_ref>
</app_version>
</app_info>

Also, I'm running (BOINC Manager Version: 7.0.28 (x64) )

[KieX]

Originally Posted by Mindweaver

I put the file in and I'm crunching 2x GPU WU's but no CPU WU's? wtf? I reset the project.. but still no CPU work...

10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app sn2s platform windows_intelx86 ver 620 class ; discarding SN2S_AAG25601_0000168_0238_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app sn2s platform windows_intelx86 ver 620 class ; discarding SN2S_AAG25601_0000168_0271_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app sn2s platform windows_intelx86 ver 620 class ; discarding SN2S_AAG25601_0000168_0264_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app dsfl platform windows_intelx86 ver 625 class ; discarding DSFL_00020-47_0000036_0339_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app dsfl platform windows_intelx86 ver 625 class ; discarding DSFL_00020-47_0000036_0801_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app hfcc platform windows_intelx86 ver 640 class ; discarding HFCC_target-10_00782401_target-10_0001_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app hfcc platform windows_intelx86 ver 640 class ; discarding HFCC_target-10_00782626_target-10_0001_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app hfcc platform windows_intelx86 ver 640 class ; discarding HFCC_target-10_00785022_target-10_0000_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app hfcc platform windows_intelx86 ver 640 class ; discarding HFCC_target-10_00785299_target-10_0001_0
10/23/2012 4:26:47 PM | World Community Grid | [error] No app version found for app hpf2 platform windows_intelx86 ver 640 class ; discarding qi388_00116_18

In order to receive CPU work, the CPU app needs to be explicitly stated in the app_info file. Your errors look like BOINC can't find the information to run the WU's on the server.

What does your app_info look like? Looking at it now