I'm using Boinc to let my HD5970 do some crunching. The CPU is crunching on the Muon-project from Stephen Brooks.

The problem is that Boinc spawns 8 children of the MW-CPU app, but as the CPU is utilized already, they do not make much progress and only eat memory.

How to disable the CPU completely (besides the GPU supporting tasks) and let only the GPU work ?

To turn the cpu crunching off: My Account->MilkyWay@home preferences->uncheck CPU. It will still use a small portion to run the GPU tasks.

---

Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.

Aaaah! Who would think one has to configure his computer via the web browser ;-)

Thanks!

Now the client is also not waiting for the CPU tasks to finish before requesting new WUs.

You are, in fact, not configuring your computer via the web browser. What you are doing is telling the server not to send you CPU tasks to crunch. It all seems quite logical to me.

---

Ah the tasks are different as well. I see.

Another problem: when windows 7 uses some Areo fades etc. my mouse is very slow (delayed) and also these fades are not smooth at all. With Folding @ Home I did not have these problems. I also had to set the following env vars:

CAL_NO_FLUSH=1
BROOK_YIELD=2
FLUSH_INTERVAL=128
CAL_PRE_FLUSH=1

http://jiggmin.com/forum/showthread.php?t=36514 wrote:

FLUSH_INTERVAL is what will affect graphics performance (the 2D lag phenomenon for example). This is the number of functions sent to the GPU in one go. The GPU will not do anything else, including refreshing the screen until processing of the batch of commands ends. A low value reduces the time F@H monopolises the GPU, and the response time of the interface increases. However, the lower the value becomes, the higher the CPU load related to the OS and the driver becomes, so there is a trade-off between the performance of F@H and the fluidity of the interface. If the batch is too large, it can cause a VPU Recover, the driver thinking that the GPU has hung (when it is just taking too long to respond).

CAL_NO_FLUSH and CAL_PRE_FLUSH change the method of submitting batches of functions to the hardware. CAL_NO_FLUSH changes how the batches of functions are built. CAL_PRE_FLUSH allows caching the batch, in order to prepare the next batch in advance while the GPU handles the current batch.

BROOK_YIELD has several modes: 0/1/2. 0 will monopolise the CPU to have the lowest latency to requests from the GPU. 1 will release the CPU while waiting for response from the GPU to process all commandsof the same or lower priority than the GPU core. 2 will release the CPU for every process, regardless of its priority. Now, for very small values of FLUSH_INTERVAL and small proteins, it is likely that the GPU is almost finished when the CPU is released. The GPU must wait to regain access to the CPU, which may take up to a millisecond. A high-end GPU will complete most such batches in less than 100 microseconds, therefore the period of waiting for access to the CPU can have a big impact on performance. With a high value of FLUSH_INTERVAL, it is easy to build several milliseconds of work, making the wait period less of an issue.

Are there any of these settings for MW? Because I do not accept the way MW uses the GPU now, but I do want to be crunching while using the desktop for surfing and some other light 2D stuff.

Thanks in advance!

There are ways to adjust the behavior of the ATI application (developed by Cluster Physik) used here. They're however not trivial.
They're several command line parameters you can adjust to configure the GPU utilization. Basically you need to use something called anonymous platform. This however means that you've to set up everything manual, e.g. if the project releases a new application you've to download it manually and edit the configuration file to enable it. That means you need to monitor the project to find out when changes are necessary. If you just want to set and forget I would strongly advise against it.

If you still want to continue:
1) Download Cluster Physiks application package
here
2) Unzip it to a temporary location
3) read the included readme.txt
4) Quit BOINC
5) Copy everything from the Win64_Ati-subfolder to your milkyway directory (probably C:\ProgramData\BOINC\projects\milkyway.cs.rpi.edu_milkyway)
6) edit the app_info.xml file to set your options. As a starting point the content of mine:

<app_info>
 <app>
  <name>milkyway</name>
  <user_friendly_name>Milkyway@Home</user_friendly_name>
 </app>
 <file_info>
  <name>astronomy_0.20b_ATI_x64_ati.exe</name>
  <executable/>
 </file_info>
 <file_info>
  <name>brook64.dll</name>
  <executable/>
 </file_info>
 <app_version>
  <app_name>milkyway</app_name>
  <version_num>21</version_num>
    <flops>1.0e11</flops>
    <avg_ncpus>0.05</avg_ncpus>
    <max_ncpus>1</max_ncpus>
    <coproc>
      <type>ATI</type>
      <count>1</count>
    </coproc>
    <plan_class>ati13ati</plan_class>
    <cmdline>w1.3</cmdline>
  <file_ref>
   <file_name>astronomy_0.20b_ATI_x64_ati.exe</file_name>
   <main_program/>
  </file_ref>
  <file_ref>
   <file_name>brook64.dll</file_name>
  </file_ref>
 </app_version>
</app_info>

This is the part you need no change to keep the system more responsive. Your system should become more responsive when you raise the value (standard is: w1.0):
<cmdline>w1.3</cmdline>
Everytime you want to change the value you have to stop BOINC, change it and start BOINC again.

7) Start BOINC

---

Starfire

Thanks a lot for this clear explanation and not referring me to the search! I did search the forum and found some posts about this xml, but I could not find it on my system ;-)

Now I know why ;-)

I will experiment with this. Isn't it better to include this setting with the Boinc Manager or so?

So i need to create an anon. platform. But does this also mean that the stats and points are not accredited to my account?

No, they are still credited to you. It is just listed as anonymous because of the app_info.xml file.

If the server know what app you are running it says so, if it does not it lists it as anonymous.

---