If I take the brackets off (what someone else gave me too) then I get the red error message saying it can't parse exclude command.

Can you please show the error message for that from your event log?

David Anderson writes:
[...] is used in the documentation to indicate that an element is optional.

---

I'll post in multiple replies so its a little easier to follow:

Here is the cc_config.xml file:
<cc_config>
<log_flags>
</log_flags>
<options>
<client_version_check_url>http://www.worldcommunitygrid.org/download.php?xml=1</client_version_check_url>
<client_download_url>http://www.worldcommunitygrid.org/download.php</client_download_url>
<network_test_url>http://www.ibm.com/</network_test_url>
<start_delay>10</start_delay>
<max_file_xfers>16</max_file_xfers>
<max_file_xfers_per_project>16</max_file_xfers_per_project>

<exclude_gpu>
<url>http://www.worldcommunitygrid.org</url>
[<type>nvidia</type>]
</exclude_gpu>


</options>
</cc_config>



Here is the log when I use brackets:

2/8/2013 1:06:55 AM | | Starting BOINC client version 7.0.44 for windows_x86_64
2/8/2013 1:06:55 AM | | log flags: file_xfer, sched_ops, task
2/8/2013 1:06:55 AM | | Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6
2/8/2013 1:06:55 AM | | Data directory: C:\ProgramData\BOINC
2/8/2013 1:06:55 AM | | Running under account Jeff
2/8/2013 1:06:55 AM | | Processor: 4 GenuineIntel Intel(R) Core(TM) i5-3570K CPU @ 3.40GHz [Family 6 Model 58 Stepping 9]
2/8/2013 1:06:55 AM | | Processor: 256.00 KB cache
2/8/2013 1:06:55 AM | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 sse4_1 sse4_2 syscall nx lm vmx tm2 popcnt aes pbe
2/8/2013 1:06:55 AM | | OS: Microsoft Windows 7: Professional x64 Edition, Service Pack 1, (06.01.7601.00)
2/8/2013 1:06:55 AM | | Memory: 7.97 GB physical, 15.93 GB virtual
2/8/2013 1:06:55 AM | | Disk: 167.58 GB total, 52.13 GB free
2/8/2013 1:06:55 AM | | Local time is UTC -5 hours
2/8/2013 1:06:55 AM | | CUDA: NVIDIA GPU 0: GeForce GTX 670 (driver version 310.90, CUDA version 5.0, compute capability 3.0, 2048MB, 1830MB available, 2634 GFLOPS peak)
2/8/2013 1:06:55 AM | | CAL: ATI GPU 0: AMD Radeon HD 7700 series (Cape Verde) (CAL version 1.4.1741, 1024MB, 984MB available, 3840 GFLOPS peak)
2/8/2013 1:06:55 AM | | OpenCL: NVIDIA GPU 0: GeForce GTX 670 (driver version 310.90, device version OpenCL 1.1 CUDA, 2048MB, 1830MB available, 2634 GFLOPS peak)
2/8/2013 1:06:55 AM | | OpenCL: ATI GPU 0: AMD Radeon HD 7700 series (Cape Verde) (driver version 1084.4 (VM), device version OpenCL 1.2 AMD-APP (1084.4), 1024MB, 984MB available, 3840 GFLOPS peak)
2/8/2013 1:06:55 AM | Asteroids@home | Found app_config.xml
2/8/2013 1:06:55 AM | World Community Grid | Found app_config.xml
2/8/2013 1:06:55 AM | World Community Grid | Config: excluded GPU. Type: nvidia. App: all. Device: all
2/8/2013 1:06:55 AM | Asteroids@home | URL http://asteroidsathome.net/boinc/; Computer ID 9004; resource share 100
2/8/2013 1:06:55 AM | WUProp@Home | URL http://wuprop.boinc-af.org/; Computer ID 47366; resource share 1
2/8/2013 1:06:55 AM | GPUGRID | URL http://www.gpugrid.net/; Computer ID 128329; resource share 100
2/8/2013 1:06:55 AM | World Community Grid | URL http://www.worldcommunitygrid.org/; Computer ID 2074207; resource share 200
2/8/2013 1:06:55 AM | World Community Grid | General prefs: from World Community Grid (last modified 04-Feb-2013 03:05:10)
2/8/2013 1:06:55 AM | World Community Grid | Computer location: home



Here it is asking for nvidia:
2/8/2013 1:08:13 AM | World Community Grid | Requesting new tasks for NVIDIA and ATI


Here it is working nvidia:
2/8/2013 1:14:07 AM | World Community Grid | Starting task X0930107161378200903111110_0 using hcc1 version 705 (nvidia_hcc1) in slot 10
2/8/2013 1:14:07 AM | World Community Grid | Starting task X0960109170138200904241241_1 using hcc1 version 705 (nvidia_hcc1) in slot 11


I'll post no brackets in next message.

Here is the cc_config.xml with no brackets:
<cc_config>
<log_flags>
</log_flags>
<options>
<client_version_check_url>http://www.worldcommunitygrid.org/download.php?xml=1</client_version_check_url>
<client_download_url>http://www.worldcommunitygrid.org/download.php</client_download_url>
<network_test_url>http://www.ibm.com/</network_test_url>
<start_delay>10</start_delay>
<max_file_xfers>16</max_file_xfers>
<max_file_xfers_per_project>16</max_file_xfers_per_project>

<exclude_gpu>
<url>http://www.worldcommunitygrid.org</url>
<type>nvidia</type>
</exclude_gpu>


</options>
</cc_config>


Ok, I tried the above config and didn't get an error message about nvidia this time. Apparently reinstalling .44 fixed that part. Its still requesting nvidia task and running them. Any time frame on a fix for this?

Its still requesting nvidia task and running them. Any time frame on a fix for this?

You'll have to show this in a log. And then one with log_flags <cpu_sched_debug> and <work_fetch_debug> on.

Like so:

<cc_config>
<log_flags>
<cpu_sched_debug>1</cpu_sched_debug>
<work_fetch_debug>1</work_fetch_debug>
</log_flags>
<options>
<client_version_check_url>http://www.worldcommunitygrid.org/download.php?xml=1</client_version_check_url>
<client_download_url>http://www.worldcommunitygrid.org/download.php</client_download_url>
<network_test_url>http://www.ibm.com/</network_test_url>
<start_delay>10</start_delay>
<max_file_xfers>16</max_file_xfers>
<max_file_xfers_per_project>16</max_file_xfers_per_project>
<exclude_gpu>
<url>http://www.worldcommunitygrid.org</url>
<type>nvidia</type>
</exclude_gpu>
</options>
</cc_config>

If you really mean to be using a beta version, you should update to 7.0.48 as well. Download links available in the change log thread.

---

<app_config>
<app>
<name>astropulse_v6</name>
<max_concurrent>2</max_concurrent>
<gpu_versions>
<gpu_usage>.5</gpu_usage>
<cpu_usage>.997</cpu_usage>
</gpu_versions>
</app>
</app_config>

This only runs 2 instances on 1 device
How do I make it run 2 instances on all 3 devices?

You need to set the gpu option <use_all_gpus>1</use_all_gpus> in the cc_config.xml to use all gpus, then increase the <max_concurrent> to 6. Theoretically, the <gpu_usage> limits the per-card to 2 [try and see]. Of course the CPU must be able to handle all 6. At .997 you're kind of saying to use 6 CPU threads.

---

Coelum Non Animum Mutant, Qui Trans Mare Currunt

You need to set the gpu option <use_all_gpus>1</use_all_gpus> in the cc_config.xml to use all gpus, then increase the <max_concurrent> to 6. Theoretically, the <gpu_usage> limits the per-card to 2 [try and see]. Of course the CPU must be able to handle all 6. At .997 you're kind of saying to use 6 CPU threads.

Thank you for confirming this.
The use all gpu's doesn't seems to be needed anymore.
I increased max concurrent to 4 and then it used 2 gpu's
I will now increase it to 6
Thank you again.

-=sometimes math isn't logic to me=-

The use all gpu's doesn't seems to be needed anymore.

[I] dare not say if <use_all_gpus>1</use_all_gpus> is required under the newer clients if the capability of cards is different. Just know that if equal, they're all used. Maybe someone positively in the know can write an expansion on this entry in this wiki as clarification: http://boinc.berkeley.edu/wiki/Client_configuration#Application_configuration . It does not harm to make the entry, then use any of the next 3 <options> to disable a specific GPU if need be.

<ignore_ati_dev>N</ignore_ati_dev>
<ignore_cuda_dev>N</ignore_cuda_dev>
<ignore_nvidia_dev>N</ignore_nvidia_dev>

But, since I don't have a device with 2 or 3 GPU's it's from reading the manual and any reports of practical experience... Don't know if the actual functionality of any of these GPU related options changed between versions. The manuals always seem to be aimed at describing the latest client [whatever latest is].

---

Coelum Non Animum Mutant, Qui Trans Mare Currunt

The use all gpu's doesn't seems to be needed anymore.

[I] dare not say if <use_all_gpus>1</use_all_gpus> is required under the newer clients if the capability of cards is different. Just know that if equal, they're all used. Maybe someone positively in the know can write an expansion on this entry in this wiki as clarification: http://boinc.berkeley.edu/wiki/Client_configuration#Application_configuration . It does not harm to make the entry, then use any of the next 3 <options> to disable a specific GPU if need be.

<ignore_ati_dev>N</ignore_ati_dev>
<ignore_cuda_dev>N</ignore_cuda_dev>
<ignore_nvidia_dev>N</ignore_nvidia_dev>

But, since I don't have a device with 2 or 3 GPU's it's from reading the manual and any reports of practical experience... Don't know if the actual functionality of any of these GPU related options changed between versions. The manuals always seem to be aimed at describing the latest client [whatever latest is].

I saw a post in some other thread about it and deleted it BM7.0.28 and it seemed to work so I don't use it.
But I have 1 5970 with 2 GPU's and 1 5850 so the cards are close family.

Thanks soooo much for this thread!

I am now running 10 gpu tasks and four regular tasks.:)
And I thought I would see a time increase in the gpu jobs but not so far.

this is what I used:

<app_config>
<app>
<name>astropulse_v6</name>
<max_concurrent>2</max_concurrent>
<gpu_versions>
<gpu_usage>.1</gpu_usage>
<cpu_usage>.05</cpu_usage>
</gpu_versions>
</app>
</app_config>


The .05 gives enough room to still run the 4 cpu jobs for me. but might make that lower seeing as the computer is lagging bad. lol

Thanks again

Ok just found a wrinkle in this system. LOL

Open Malaria has two apps:

openMalariaA
openMalariaB

So How do I tell BOINC I only want 1 OM running at any time?
Not 1 from A and 1 from B LOL

Thanks again.

BTW while running 10 GPU projects was fun went back to one for now the end times were to long.

You can set in Malaria project preferences which applications you want to run.
Or set a very low minimum cache and minimum additional. Zero and zero will get you 1 task per CPU core.

---