Hello,
Are we gong to see a GPU utilisation factor in the stock SAH settings anytime?
Thank you -

---

Are you referring more to resource scheduling type scenarios ? or more toward direct control of GPU usage in the application sense ?

Not sure about the AMD GPU app side, as your listed PC seems to have, though I've certainly been looking at ideas for future more configurable/controllable nVidia Cuda running in a number of possible user scenarios.

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

I expect I mean direct control of GPU usage i.e. running two or three WUs per GPU. The app_info.xml file that can produce this kiind of activity in my system is tough to re-author every time the stock sci-apps change; it would be nice to see a "How many do you want to try to run per GPU" option in the settings.
Thank you -

---

They've introduced with newer Boinc clients an app_config.xml to allow control of the number of instances on a per application basis, with/without an app_info.xml. I'm told that's a simpler way, that should work running stock or under anonymous platform, though I use older Boinc clients so will leave others that use it to explain the mechanism there.

I'm not aware of other work in the area of configurability, though for the third party (anonymous platform) app side, I've been examining ways to allow better/easier local configuration to per device level. That's quite a way off, & initially targeted for external tools, though backstage discussions about some areas have suggested parts of that effort might eventually be wanted back in the Boinc client as well.

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

This is what mine is set to currently. It allows me to run 2 MB or 1 MB and 1 AP on each card.

<app_config>
   <app>
      <name>setiathome_v7</name>
      <gpu_versions>
          <gpu_usage>0.49</gpu_usage>
          <cpu_usage>.04</cpu_usage>
      </gpu_versions>
    </app>
	<app>
        <name>astropulse_v6</name>
		 <gpu_versions>
          <gpu_usage>0.51</gpu_usage>
          <cpu_usage>1</cpu_usage>
      </gpu_versions>
    </app>
</app_config>

It also reserves a full CPU for the AP work.

---

Great! So it does.
Thanks -

---

Great! So it does.
Thanks -

It looks like all your work is coming in with about an average of 90 seconds from the server and you are erroring out everything right now.

---

Hows about your CPU ?
Did you reserve a core ?
Running 2 GPU`s on a dual core CPU is difficult.

---

With each crime and every kindness we birth our future.

Yeas, everything was erroring! I was using the 4.01 Lunatics apps with no problems until a few days ago on this system*. Followed the advice given out for similar issues, and reinstalled the 4.01 but no joy. So I'm now back on the stock apps which are 'dueling' OK.

*Windows 7 64 bit.
Opteron 185 + 2x HD6950.
185 has 1 reserved core.

---

Yeas, everything was erroring! I was using the 4.01 Lunatics apps with no problems until a few days ago on this system*. Followed the advice given out for similar issues, and reinstalled the 4.01 but no joy. So I'm now back on the stock apps which are 'dueling' OK.

*Windows 7 64 bit.
Opteron 185 + 2x HD6950.
185 has 1 reserved core.

With that setup you'd be better off not using your CPU, keep both its cores free to feed those cards.

Cheers.