it should be in the right project order of poem "gpugrid" two orders up from your order where cc_config os contained when i remember right.

---

DSKAG Austria Research Team: http://www.research.dskag.at

My BOINC data folder is a little different because I have my main windows installation on an SSD and all data-related things on a hard disk, but the directory is something like this for me:

D:\HDDProgramFiles\BOINC Data\projects\gpugrid\app_info.xml

What you need to do is the following:
Since BOINC discards all tasks the first time it finds an app_info, you need to finish your current ones (assuming you don't want to just abort them). To do this:
1. Open boinc, click on GPUgrid in the "projects" tab and click "no new tasks"
2. Wait for all current tasks to end
3. Exit BOINC

Then, you need to make the app_info file. So:
1. Open notepad
2. Copy and paste the contents from my first post into notepad
3. Save it as app_info.xml - with a .xml extension. Notepad by default saves with a .txt extension. Worst case, if you save it as .txt you can just rename the file, remove the ".txt" and make it ".xml".

Save the file within the project folder.

In a default install, the BOINC data folder will be:

Windows 98/SE/ME: C:\Windows\All Users\BOINC\ or C:\Windows\Profiles\All Users\BOINC\ (*)
Windows 2000/XP: C:\Documents and Settings\All Users\Application Data\BOINC\ (*)
Windows Vista/Windows 7: C:\ProgramData\BOINC\ (*)
Linux: wherever you unpack it/BOINC/
Macintosh OS X: /Library/Applications Support/BOINC/

(*) This directory may well be hidden, so either put the path to it directly into Windows Explorer, or instruct Windows Explorer to show hidden files and folders.

Once you open the BOINC data folder, you'll find a folder called "projects". Open that, and you should have a folder for every BOINC project running on that PC. Open the gpugrid folder, and put the app_info.xml in there. Then start up BOINC, set it to allow new tasks, and look at the notices and task list to check that it was a success.

Luke

Where is the file app_info.xml in the ProgramData directory from which GPUGRD
program works?

I know it does not work in the cc_config file in Bonic program.

So in other words, where are you or other people putting the program to have 2 wu's work at once.

Thanks
spout23

My previous message was aimed at someone who didn't know where the BOINC data folder was. Since you already know where the cc_config is, then the short answer for you is to open the "projects" folder, which is in the same directory as cc_config, then open the "gpugrid" folder.

In other words: BOINC_data_folder\projects\gpugrid\app_info.xml

Well what can I say, the file does work in the place you said.
But I would recommend that on the most expert people among us in the
Grid do this procedure. I'm a superuser on the computer,
But I'm Not a expert.

Thanks for your Help.

It appears there are 2 app versions being distributed on the LONG queue ... 615 and 616. I added a new app_version section for 616 and it seems to be working fine.

---

Thanks - Steve

It appears there are 2 app versions being distributed on the LONG queue ... 615 and 616. I added a new app_version section for 616 and it seems to be working fine.

Hmm. So you duplicated the entire section between <app_version> and </app_version> ? Could you post the complete app_info?

I always wonder whether it would be possible to have separate sections of the app_info.xml for Nathan's and Paola's so that we could run only one nathan and two paolas automatically. Any idea how this could be done?

The app_info only knows about the application, it has no idea of what types of WUs it is running.
I've now started running this on my i7-920 Win7 x64 in conjunction with CPU tasks from other projects ... I left a more core free than necessarey and something odd is happening ... instead of splitting the core in half for the 2 instances of ACEMD it is using a bit more than 1/2 for each and I'm currently running at a pace to complete 2 in 16 hours total!!! I have no idea what's going on but I'm not going to question too much.

Here is the app_info I am currently using. A couple of points to note ... I upped the MEM by simply changing the leading 1 to a 2 which is not exact but close enough and I have also upped the avg num cpus. I was using this same file on my i7-980 Win7 x64 rig with the 660Ti and while I was getting better efficiency overall it took 30 hours to complete 2 WUs.

This morning after discarding quite a few NATE WUs to capture more PAOLOAs for the 660, BOINC started saying I had a bad app_info??? Whatever ... it's working great on the 670 so I'm going to leave it alone.

<app_info>
<app>
<name>acemdlong</name>
<user_friendly_name>Long runs (8-12 hours on fastest card)</user_friendly_name>
</app>
<file_info>
<name>acemd.2562.cuda42</name>
<executable/>
</file_info>
<file_info>
<name>cudart32_42_9.dll</name>
<executable/>
</file_info>
<file_info>
<name>cufft32_42_9.dll</name>
<executable/>
</file_info>
<file_info>
<name>tcl85.dll</name>
<executable/>
</file_info>
<app_version>
<app_name>acemdlong</app_name>
<version_num>615</version_num>
<platform>windows_intelx86</platform>
<avg_ncpus>0.50</avg_ncpus>
<max_ncpus>1.000000</max_ncpus>
<flops>6.0e11</flops>
<plan_class>cuda42</plan_class>
<api_version>6.7.0</api_version>
<coproc>
<type>CUDA</type>
<count>.5</count>
</coproc>
<gpu_ram>2073741824.00000</gpu_ram>
<file_ref>
<file_name>acemd.2562.cuda42</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_42_9.dll</file_name>
<open_name>cudart32_42_9.dll</open_name>
<copy_file/>
</file_ref>
<file_ref>
<file_name>cufft32_42_9.dll</file_name>
<open_name>cufft32_42_9.dll</open_name>
<copy_file/>
</file_ref>
<file_ref>
<file_name>tcl85.dll</file_name>
<open_name>tcl85.dll</open_name>
<copy_file/>
</file_ref>
</app_version>
<app_version>
<app_name>acemdlong</app_name>
<version_num>616</version_num>
<platform>windows_intelx86</platform>
<avg_ncpus>0.50</avg_ncpus>
<max_ncpus>1.000000</max_ncpus>
<flops>6.0e11</flops>
<plan_class>cuda42</plan_class>
<api_version>6.7.0</api_version>
<coproc>
<type>CUDA</type>
<count>.5</count>
</coproc>
<gpu_ram>2073741824.00000</gpu_ram>
<file_ref>
<file_name>acemd.2562.cuda42</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_42_9.dll</file_name>
<open_name>cudart32_42_9.dll</open_name>
<copy_file/>
</file_ref>
<file_ref>
<file_name>cufft32_42_9.dll</file_name>
<open_name>cufft32_42_9.dll</open_name>
<copy_file/>
</file_ref>
<file_ref>
<file_name>tcl85.dll</file_name>
<open_name>tcl85.dll</open_name>
<copy_file/>
</file_ref>
</app_version>
</app_info>

---

Thanks - Steve

So ... I figured out what the "bad app_info" message really is (I should read the message closer) ... it really was telling me that I had already puled 18 WUs for the day, so this means I won;t get any more. I just had this happen on my 670 so I'm giving up this venture and going back to crunching 1 at a time :sad_panda:

---

Thanks - Steve

I'm receiving these PAOLA_3EKO tasks less and less frequently lately, so this problem will be gone soon.
I hope the upcoming long tasks won't have similar problems.

I'm receiving these PAOLA_3EKO tasks less and less frequently lately, so this problem will be gone soon.
I hope the upcoming long tasks won't have similar problems.

I've gotten 2 additional PAOLA's with different names acting like the 3EKO, it looks as though there are 2 more runs coming out of the gate.

---

The app_info only knows about the application, it has no idea of what types of WUs it is running.
I've now started running this on my i7-920 Win7 x64 in conjunction with CPU tasks from other projects ... I left a more core free than necessarey and something odd is happening ... instead of splitting the core in half for the 2 instances of ACEMD it is using a bit more than 1/2 for each and I'm currently running at a pace to complete 2 in 16 hours total!!! I have no idea what's going on but I'm not going to question too much.

Here is the app_info I am currently using. A couple of points to note ... I upped the MEM by simply changing the leading 1 to a 2 which is not exact but close enough and I have also upped the avg num cpus. I was using this same file on my i7-980 Win7 x64 rig with the 660Ti and while I was getting better efficiency overall it took 30 hours to complete 2 WUs.

This morning after discarding quite a few NATE WUs to capture more PAOLOAs for the 660, BOINC started saying I had a bad app_info??? Whatever ... it's working great on the 670 so I'm going to leave it alone.

<app_info>
<app>
<name>acemdlong</name>
<user_friendly_name>Long runs (8-12 hours on fastest card)</user_friendly_name>
</app>
<file_info>
<name>acemd.2562.cuda42</name>
<executable/>
</file_info>
<file_info>
<name>cudart32_42_9.dll</name>
<executable/>
</file_info>
<file_info>
<name>cufft32_42_9.dll</name>
<executable/>
</file_info>
<file_info>
<name>tcl85.dll</name>
<executable/>
</file_info>
<app_version>
<app_name>acemdlong</app_name>
<version_num>615</version_num>
<platform>windows_intelx86</platform>
<avg_ncpus>0.50</avg_ncpus>
<max_ncpus>1.000000</max_ncpus>
<flops>6.0e11</flops>
<plan_class>cuda42</plan_class>
<api_version>6.7.0</api_version>
<coproc>
<type>CUDA</type>
<count>.5</count>
</coproc>
<gpu_ram>2073741824.00000</gpu_ram>
<file_ref>
<file_name>acemd.2562.cuda42</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_42_9.dll</file_name>
<open_name>cudart32_42_9.dll</open_name>
<copy_file/>
</file_ref>
<file_ref>
<file_name>cufft32_42_9.dll</file_name>
<open_name>cufft32_42_9.dll</open_name>
<copy_file/>
</file_ref>
<file_ref>
<file_name>tcl85.dll</file_name>
<open_name>tcl85.dll</open_name>
<copy_file/>
</file_ref>
</app_version>
<app_version>
<app_name>acemdlong</app_name>
<version_num>616</version_num>
<platform>windows_intelx86</platform>
<avg_ncpus>0.50</avg_ncpus>
<max_ncpus>1.000000</max_ncpus>
<flops>6.0e11</flops>
<plan_class>cuda42</plan_class>
<api_version>6.7.0</api_version>
<coproc>
<type>CUDA</type>
<count>.5</count>
</coproc>
<gpu_ram>2073741824.00000</gpu_ram>
<file_ref>
<file_name>acemd.2562.cuda42</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart32_42_9.dll</file_name>
<open_name>cudart32_42_9.dll</open_name>
<copy_file/>
</file_ref>
<file_ref>
<file_name>cufft32_42_9.dll</file_name>
<open_name>cufft32_42_9.dll</open_name>
<copy_file/>
</file_ref>
<file_ref>
<file_name>tcl85.dll</file_name>
<open_name>tcl85.dll</open_name>
<copy_file/>
</file_ref>
</app_version>
</app_info>

Thanks for that. Today I finally reached #1 spot in the Malta leaderboard (using a more time-profitable project), so as of now I have once again started running GPUGRID. I'm using your app_info - your idea of changing the avg_ncpus was a good idea - BOINC now runs two GPUgrid tasks and 7 CPU tasks at once on my computer (before it used to try to run 8 CPU tasks but one of them was always running slower than the other 7 - likely due to the way windows schedules CPU time between tasks when it has an excess of active threads...).

I'm currently running a Nathan and a Paola simultaneously. The GPU load is 97-99% (a record for me). Computer is just as responsive as usual, which is good. GPU RAM usage is 1700MB, which is also a record as it's usually under 700MB with other BOINC GPU projects.

I'll try and remember to report back when the tasks complete with the completion time and calculated rate of points. Then we can compare the points per second of running two at once with the points per second of running just one Nathan alone.