Working app_info files

[Mydog]

Keep getting a lot of errors on the 6990+6970 rig, twelve pages in 24 hours.

(hours) Claimed/ Granted BOINC Credit
X0930074500355200609121543_ 1-- Cruncher1-PC Error 11/6/12 17:51:58 11/6/12 23:13:46 0.00 / 0.01 0.8 / 0.0
X0930074500341200609121543_ 0-- Cruncher1-PC Error 11/6/12 17:51:58 11/6/12 23:13:46 0.00 / 0.00 0.7 / 0.0
X0930073401252200608181020_ 1-- Cruncher1-PC Error 11/6/12 17:50:46 11/6/12 22:55:09 0.00 / 0.01 0.8 / 0.0
X0930073400602200608181031_ 0-- Cruncher1-PC Error 11/6/12 17:49:28 11/6/12 22:49:21 0.00 / 0.01 1.4 / 0.0
X0960073371108200608241458_ 0-- Cruncher1-PC Error 11/6/12 17:47:16 11/6/12 22:36:58 0.00 / 0.01 1.2 / 0.0
X0960073371116200608241458_ 0-- Cruncher1-PC Error 11/6/12 17:47:16 11/6/12 22:36:58 0.00 / 0.01 1.1 / 0.0
X0960073371111200608241458_ 1-- Cruncher1-PC Error 11/6/12 17:47:16 11/6/12 22:32:14 0.00 / 0.01 1.1 / 0.0
X0930074550973200609261742_ 1-- Cruncher1-PC Error 11/6/12 17:03:14 11/6/12 22:20:29 0.00 / 0.01 1.0 / 0.0
X0900074561124200608301609_ 1-- Cruncher1-PC Error 11/6/12 17:00:24 11/6/12 22:20:29 0.00 / 0.00 0.6 / 0.0
X0900074051353200609182008_ 0-- Cruncher1-PC Error 11/6/12 16:58:04 11/6/12 22:20:29 0.00 / 0.01 1.0 / 0.0
X0900074050917200609182014_ 1-- Cruncher1-PC Error 11/6/12 16:57:06 11/6/12 22:20:29 0.00 / 0.01 1.0 / 0.0
X0960074550058200609051831_ 1-- Cruncher1-PC Error 11/6/12 16:49:22 11/6/12 22:20:29 0.00 / 0.01 0.9 / 0.0
X0960074550060200609051831_ 0-- Cruncher1-PC Error 11/6/12 16:49:22 11/6/12 22:20:29 0.00 / 0.01 0.8 / 0.0
X0930074040911200609181949_ 1-- Cruncher1-PC Error 11/6/12 16:44:57 11/6/12 22:20:29 0.00 / 0.00 0.7 / 0.0
X0930074040202200609182001_ 1-- Cruncher1-PC Error 11/6/12 16:43:39 11/6/12 22:20:29 0.00 / 0.01 1.2 / 0.0

Most of them with this on them:

<core_client_version>7.0.28</core_client_version>
<![CDATA[
<message>
No process is on the other end of the pipe. (0xe9) - exit code 233 (0xe9)
</message>
<stderr_txt>

</stderr_txt>
]]> 

and a few with this:

<core_client_version>7.0.28</core_client_version>
<![CDATA[
<message>
No process is on the other end of the pipe. (0xe9) - exit code 233 (0xe9)
</message>
<stderr_txt>
Commandline: projects/www.worldcommunitygrid.org/wcg_hcc1_img_6.56_windows_intelx86__ati_hcc1 X0930074500341200609121543.jp2 --device 1 
<app_init_data> 
<major_version>7</major_version> 
<minor_version>0</minor_version> 
<release>28</release> 
<app_version>656</app_version> 
<app_name>hcc1</app_name> 
<project_preferences> 


<color_scheme>Tahiti Sunset</color_scheme> 
<max_frames_sec>2</max_frames_sec> 
<max_gfx_cpu_pct>1.0</max_gfx_cpu_pct> 
</project_preferences> 

<project_dir>C:\ProgramData\BOINC/projects/www.worldcommunitygrid.org</project_dir>
<boinc_dir>C:\ProgramData\BOINC</boinc_dir> 
<wu_name>X0930074500341200609121543</wu_name> 
<result_name>X0930074500341200609121543_0</result_name> 
<comm_obj_name>boinc_1</comm_obj_name> 
<slot>14</slot> 
<wu_cpu_time>0.000000</wu_cpu_time> 
<starting_elapsed_time>0.000000</starting_elapsed_time> 
<using_sandbox>0</using_sandbox> 
<user_total_credit>30711497.815184</user_total_credit> 
<user_expavg_credit>179649.631246</user_expavg_credit> 
<host_total_credit>1025194.232916</host_total_credit> 
<host_expavg_credit>21706.211695</host_expavg_credit> 
<resource_share_fraction>1.000000</resource_share_fraction> 
<checkpoint_period>1200.000000</checkpoint_period> 
<fraction_done_start>0.000000</fraction_done_start> 
<fraction_done_end>1.000000</fraction_done_end> 
<gpu_type>ATI</gpu_type> 
<gpu_device_num>1</gpu_device_num> 
<gpu_opencl_dev_index>1</gpu_opencl_dev_index> 
<ncpus>1.000000</ncpus> 
<rsc_fpops_est>13223751080296.000000</rsc_fpops_est> 
<rsc_fpops_bound>264475021605920.000000</rsc_fpops_bound> 
<rsc_memory_bound>78643200.000000</rsc_memory_bound> 
<rsc_disk_bound>50000000.000000</rsc_disk_bound> 
<computation_deadline>1352820478.000000</computation_deadline> 
<vbox_window>0</vbox_window> 
</app_init_data> 
INFO: gpu_type set in init_data.xml to ATI 
INFO: gpu_device_num set in init_data.xml to 1 
Boinc requested ATI gpu device number1 
Found compute platform Advanced Micro Devices, Inc. 
Selecting this platform 
CL_DEVICE_NAME: Cayman 
CL_DEVICE_VENDOR: Advanced Micro Devices, Inc. 
CL_DEVICE_VERSION: 1016.4 (VM) 
CL_DEVICE_MAX_COMPUTE_UNITS:  
CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS: 3 
CL_DEVICE_MAX_WORK_ITEM_SIZES: 256 / 256 / 256 
CL_DEVICE_MAX_WORK_GROUP_SIZE: 256 
CL_DEVICE_MAX_CLOCK_FREQUENCY: 830 MHz 
CL_DEVICE_ADDRESS_BITS:     32 
CL_DEVICE_MAX_MEM_ALLOC_SIZE:     512 MByte 
CL_DEVICE_GLOBAL_MEM_SIZE:     2048 MByte 
CL_DEVICE_ERROR_CORRECTION_SUPPORT:    no 
CL_DEVICE_LOCAL_MEM_TYPE:     local 
CL_DEVICE_LOCAL_MEM_SIZE:     32 KByte 
CL_DEVICE_MAX_CONSTANT_BUFFER_SIZE:    64 KByte 
CL_DEVICE_QUEUE_PROPERTIES:     CL_QUEUE_PROFILING_ENABLE 
CL_DEVICE_EXTENSIONS: 
cl_khr_fp64 
cl_amd_fp64 
cl_khr_global_int32_base_atomics 
cl_khr_global_int32_extended_atomics 
cl_khr_local_int32_base_atomics 
cl_khr_local_int32_extended_atomics 
cl_khr_3d_image_writes 
cl_khr_byte_addressable_store 
cl_khr_gl_sharing 
cl_ext_atomic_counters_32 
cl_amd_device_attribute_query 
cl_amd_vec3 
cl_amd_printf 
cl_amd_media_ops 
cl_amd_popcnt 
cl_khr_d3d10_sharing 
cl_khr_dx9_media_sharing 
Estimated kernel execution time = 2.77717 [sec] 
ERROR: Kernel execution time estimate too high, exiting. 
00:09:23 (3920): called boinc_finish 

</stderr_txt>
]]> 

Any ideas on what to do?


Running 8 GPU WU's on each core(Tried 4 and 6 with mostly the same result) on an SR-2 with 12 GB HyperX at 1830MHz and CPU's at 4,4GHz(24x183)
Tried to lower CPU and memory speed, no change from what I can see.

[Snow Crash]

Well where the second message says "ERROR: Kernel execution time estimate too high, exiting." I believe you are simply running too many WUs per cpu and the gpu kernel is starving thus believing it is taking too long and has stopped responding. You might be able to get some relief by turning the Windows watchdog off via registry but at this point I would reduce the number of concurrent WUs to see if that helps.

[ChaosAD]

By running this app_info we are forced to run only hcc cpu and gpu wu? We cant run hcc gpu wu and allo other types of wu on cpu?

[my87csx481]

You can run gpu for hcc and cpu for other projects. The other projects need to be added to the app_info file though. There's a way to do it in this thread I believe but, I don't remember for sure now. Maybe others will comment.

[ChaosAD]

Thanks for your answer. if anyone could create such a file it would be great. Im used to run all projects. Why do you run only hcc?

[Sparky]

Well HCC is the only one that supports GPU, so that was the first one messed with in the app_info to get it at max efficiency. This really is the first time the app_info has been ever used, so it is all pretty new.

[Snow Crash]

It's quite a bit of work to pull all of the information necessary to run "all projects".
Take a look at post #115 of this thread ... OldChap has put togterh a file that has most of the projects, maybe you could test it and let us know if it works properly.

[my87csx481]

Thanks for your answer. if anyone could create such a file it would be great. Im used to run all projects. Why do you run only hcc?

What Sparky said....hcc gpu is the only current gpu project.

On my supported gpus, I run multiple threads of hcc gpu. On my cpu only crunchers, I run whatever else I choose to run from the WCG device profile page.

[EtaCarinae]

To avoid polluting the other post thread I continue here, which seems better appropriate.

Movieman, the directory you just displayed in your post in my system has the following path:

C:/Program Files (x86)/Boinc/


And the final directory where I have inserted the app_info.xml file is:

C:/Program Files (x86)/Boinc/projects/www.worldcommunitygrid.org/

I have no directory named Program Data . I use Windows Win7 Ultimate SP1.
I have indeed a directory named C:/Program Files/Boinc/
but in this directory you have the Boinc executable files and there is no projects subdirectory.

[dangaroos]

The c:\programdata directory is probably hidden and explorer is set to not display system/hidden files.

You can change this in explorer -> Organize-> Folder and search options -> View

edit :
Since you are adding a app_info file, I think that all current tasks will disappear and the boiinc client will download new ones. Happend to me the first time I added a app_info file so after that I set the client to "No new tasks" and let it finish the current workunits.

[Movieman]

The c:\programdata directory is probably hidden and explorer is set to not display system/hidden files.

You can change this in explorer -> Organize-> Folder and search options -> View

What he says! It's a hidden folder so listen to this guy!!!!

[my87csx481]

forum is slow as h311. is it me?

Anyway, I came home from work today and my two gpu only machines were not retrieving work. One was sitting idle (the ati one).

Anyone else having problems getting new work?

Also, does anyone know how to add gpu beta to the app info?

[jeanguy2]

Hello my87csx481,

XS Forum speed OK on the 2 rigs I operate. Note that if you have BOINg manager doing high GPU and/or CPU work it will slow down your browsing markedly.
Couls also be the cold climate... electrons flow more freely...????

[Evantaur]

something just happened, not getting new tasks

20737	World Community Grid	9.11.2012 1:39:10	Requesting new tasks for ATI	
20738	World Community Grid	9.11.2012 1:39:14	Scheduler request completed: got 2 new tasks	
20739	World Community Grid	9.11.2012 1:39:14	Resent lost task BETA_X0900075151221200610031537_0	
20740	World Community Grid	9.11.2012 1:39:14	Resent lost task BETA_X0900075151187200610031537_1	
20741	World Community Grid	9.11.2012 1:39:14	Result X0900071581053200606271024_1 is no longer usable	
20742	World Community Grid	9.11.2012 1:39:14	[error] App version returned from anonymous platform project; ignoring	
20743	World Community Grid	9.11.2012 1:39:14	[error] No app version found for app beta3 platform windows_intelx86 ver 705 class ati_hcc1; discarding BETA_X0900075151221200610031537_0	
20744	World Community Grid	9.11.2012 1:39:14	[error] No app version found for app beta3 platform windows_intelx86 ver 705 class ati_hcc1; discarding BETA_X0900075151187200610031537_1

[pfm3136]

I had to remove the app_info because of the beta work units, and it's working again.

[Snow Crash]

We are going to need to watch this closely ... if there are new executables associated with the beta that get promoted to production we are going to need to download them and it they are changing the version number so we will need to update our app_info files.

With great crunching comes great responsibility.

Is there someone who can host the new executables when it comes time so we don't all have to temporarily drop app_info and pull the new executables individually?

[kjeldoran]

I had to remove the app_info because of the beta work units, and it's working again.

Yep, remove your app_info and restart boinc. Abort the beta WUs and allow no new tasks. Also, make sure you opt out of the beta testing on the WCG site otherwise you will run into this again.

Side note, looks like we will have to update our app_info soon because the beta was for GPU and CPU HCC. Looks like version 705.

[bluestang]

Same here...

11/8/2012 7:14:07 PM | World Community Grid | Scheduler request completed: got 1 new tasks
11/8/2012 7:14:07 PM | World Community Grid | Resent lost task BETA_X0930075140273200609061426_0
11/8/2012 7:14:07 PM | World Community Grid | [error] App version returned from anonymous platform project; ignoring
11/8/2012 7:14:07 PM | World Community Grid | [error] No app version found for app beta3 platform windows_intelx86 ver 705 class ati_hcc1; discarding BETA_X0930075140273200609061426_0

[bearcatrp]

Yep, remove your app_info and restart boinc. Abort the beta WUs and allow no new tasks. Also, make sure you opt out of the beta testing on the WCG site otherwise you will run into this again.

Side note, looks like we will have to update our app_info soon because the beta was for GPU and CPU HCC. Looks like version 705.

Your sure this works? I tried but still didn't work so gave up. Uplinger stated if the beta works out, the new version will be a go, in about a week.

[my87csx481]

I ended up having to do it a couple times. Seems okay for now. Also had to untick beta for my gpu profile.

Thanks guys for the confirmation.

+1000 for gpu thread count control in Tools -> Computing Preferences

[kjeldoran]

Your sure this works? I tried but still didn't work so gave up. Uplinger stated if the beta works out, the new version will be a go, in about a week.

It worked for me as I ran into the same issue. You have to have all the beta WUs downloaded to abort them otherwise they will be resent. Once you have aborted them, you can re-add the app_info and it should be fine.

[bluestang]

Has anyone let the beta run? It definitely is different from the current!

[bearcatrp]

Thanks. It worked. Yippy!

[Bun-Bun]

Has anyone let the beta run? It definitely is different from the current!

The beta crapped out with my app_info. they got lost and I had to remove app_info in order to get WU's again. I disabled beta for now.

[kjeldoran]

Has anyone let the beta run? It definitely is different from the current!

I didn't really pay attention to the WU when I was aborting them. Are they faster? That would be nice for multiple GPU WUs .

Thanks. It worked. Yippy!

Great