thanks janusBK for the info. very good job. makes it understandable

I am sure they will be glad to have you, and it sounds like a worthwhile project too. But no one knows which one will produce the most useful science in the end (or even how you define "useful"), so picking GPU percentage is rather arbitrary. I like the fact that the IBM experts review these projects for scientific potential, and I have no idea how GPUGrid does it.

Be sure to bring your quantum computer with you.

The problem I am experiencing is extreme display refresh lag, upwards of 250 ms per refresh.

I'm fairly positive that this problem was also experienced by some of the GPUGrid.Net releases...
And they solved it by using a lower priority on the threads that the application creates/uses.
Could you please investigate if it's possible to do the same (lower the threading priority) for the WCG HCC tasks? It would be VERY much appreciated!

In the meantime, I will continue to babysit these tasks -- suspending them when I sit down at my computer, and resuming them before I walk away. It's the only way I know to still have the GPU crunch while I'm active, but not crunch WCG HCC tasks since they're far too laggy currently.

Note 1: Telling BOINC to not use GPUs while active, is not a solution. I have multiple GPUs, and don't want them all suspended!
Note 2: For your reference, as a test, I freed up all other cores on my macine (quad-core hyperthreaded so Windows sees 8 cores)... and the GPU task still causes the extreme refresh lag. So the issue is not related to CPU availability at all.
Note 3: System is Windows 8 RTM x64, Intel Core i7 965 eXtreme Edition, with 3 GPUs: GTX 460 and 2x GTS 240.

I truly believe it has to do with threading priority. PLEASE FIX IT! :)

Thanks,
Jacob

Please don't try to lecture me about my decision

If there is anyone else more committed to the success of the World Community Grid than I am I would love to shake their hand and give them the hugest of hugs

Think before you start hurling the abuse

I want the best for this project and going off half cocked is not what we have come to expect from the very best Distributed Computing Environment on the Internet... that run by IBM as the WCG.

Our [computing] resources are precious and will become ever more so in an increasingly chaotic world. We need to make the most of them.

The concept of grid computing is that the grid uses the idle portion of your pc to perform useful scientific research and as it does so it has no impact on your day to day pc activities

Since WCG's November 16th 2004 launch (note when I joined this project) this has always been the case until yesterday.

At the moment while my PC crunches 6 WCG WU's and 1 GPUGrid WU on my Nvidia card there's no impact, apart from the additional power consumption (which keeps my room warm) and the slightly faster fans, on my pc which is also rdp'ed into 2 other boxes and recording Mythbusters to my hard drive.

This is the way that Grid Computing should be and collectively I get over 100,000 BOINC points equivalent to 700,000 WCG points a day from just 4 PC's at the moment

I only wish I could provide all that capability solely to the WCG and more.

Currently I cannot

Dave

David, see if this will help.

<app_info>
<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>1</count>
<cmdline>--gpu-target-frequency 30 --gpu-polling-mode 8 --process-priority 1</cmdline>
</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_info>

Copy and paste this into a text file. Rename the file app_info.xml
Shut down boinc and place this in the WCG project folder. This should throttle down your GPU and make your desktop more usable.
The numbers 16 and 4 in the cmdline string are the ones to play with. I don't remember for sure but I think raising them makes things slower and lowering them makes things faster. If you need to tweak them shut down boinc, right click on the app_info.xml and open it with notepad. Make changes close and save then restart boinc. If you have any tasks in cache when you first put this file in the project folder you may lose them but boinc should download new ones when you restart it. If you want to tweak it leave the file in the folder while you work on it. Tasks in cache won't be affected.
I used this approach on Milkyway because it really bogged down my computer. This will increase compute times for the GPU tasks
Let me know if it works.