Hi,
I've upgraded my second system and MW has changed his main application. With the upgrade some problems came up:
- the onboard GPU is slower, uses 512MB ram, is SP
- the second GPU is a HD5830, has 1GB ram and DP.

I like to run SETI, Albert and MW on this system, but ...

How can I assign that MW is running only on the DP GPU?
How can I assign to run 2 wu's (SETI and Albert) on the HD5830 only, because both project use the gpu at ~50% only. The app_info supplies a way to run 2 wu's together, but I cannot assign to do that on the better GPU only and run single wu's on the APU.

For shure, I plan to update this system later this year, but I do not want to run the gpu (more or less) idle. And the gap will increase.

You need:
- a cc_config.xml file;
- BOINC 7.0.3 or higher. (And really, later versions do have more bug fixes, so why stay on .3?)
- the <exclude_gpu> option.
- Know which GPU is which device number. You can see this at BOINC start-up, in the event log messages.

Then, assuming that the HD5830 is device 1, your HD6550 is device 0:

<cc_config>
  <options>
    <exclude_gpu>
       <url>http://milkyway.cs.rpi.edu/milkyway/</url>
         <device_num>0</device_num>
           <type>ATI</type>
             <app>milkyway</app>
    </exclude_gpu>
    <exclude_gpu>
       <url>http://albert.phys.uwm.edu/</url>
         <device_num>1</device_num>
           <type>ATI</type>
             <app>einsteinbinary_BRP4</app>
    </exclude_gpu>
  </options>
</cc_config>

This will make sure that the HD6550 is not used by Milkyway and the HD5830 is not used by Albert. It does not affect the Nvidia GPU.

I must add, that I don't know whether the "<app>milkyway</app>" part is correct, you may want to check that. I don't have any entry for that in my client_state.xml file, so it's a guess as to what the app is called.

---

Hi Jord,

thx for your quick reply.

08.02.2012 22:17:26 | | Starting BOINC client version 7.0.12 for windows_x86_64
08.02.2012 22:17:26 | | log flags: file_xfer, sched_ops, task
08.02.2012 22:17:26 | | Libraries: libcurl/7.21.6 OpenSSL/1.0.0d zlib/1.2.5
08.02.2012 22:17:26 | | Data directory: C:\ProgramData\BOINC
08.02.2012 22:17:26 | | Running under account alex
08.02.2012 22:17:26 | | Processor: 7 AuthenticAMD AMD A8-3870 APU with Radeon(tm) HD Graphics [Family 18 Model 1 Stepping 0]
08.02.2012 22:17:26 | | Processor: 1.00 MB cache
08.02.2012 22:17:26 | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 htt pni cx16 syscall nx lm svm sse4a osvw ibs skinit wdt page1gb rdtscp 3dnowext 3dnow
08.02.2012 22:17:26 | | OS: Microsoft Windows 7: Home Premium x64 Edition, Service Pack 1, (06.01.7601.00)
08.02.2012 22:17:26 | | Memory: 7.50 GB physical, 14.99 GB virtual
08.02.2012 22:17:26 | | Disk: 111.69 GB total, 62.35 GB free
08.02.2012 22:17:26 | | Local time is UTC +1 hours
08.02.2012 22:17:26 | | VirtualBox version: 4.1.8
08.02.2012 22:17:26 | | NVIDIA GPU 0: GeForce GTX 550 Ti (driver version 285.62, CUDA version 4.10, compute capability 2.1, 1024MB, 968MB available, 705 GFLOPS peak)
08.02.2012 22:17:26 | | ATI GPU 0: ATI Radeon HD 5800 series (Cypress) (CAL version 1.4.1664, 1024MB, 991MB available, 2374 GFLOPS peak)
08.02.2012 22:17:26 | | ATI GPU 1: AMD SUPERSUMO (CAL version 1.4.1664, 512MB, 479MB available, 848 GFLOPS peak)
08.02.2012 22:17:26 | | OpenCL: NVIDIA GPU 0: GeForce GTX 550 Ti (driver version 285.62, device version OpenCL 1.1 CUDA, 1024MB, 968MB available)
08.02.2012 22:17:26 | | OpenCL: ATI GPU 0: Cypress (driver version CAL 1.4.1664 (VM), device version OpenCL 1.1 AMD-APP (851.4), 1024MB, 991MB available)
08.02.2012 22:17:26 | | OpenCL: ATI GPU 1: BeaverCreek (driver version CAL 1.4.1664 (VM), device version OpenCL 1.1 AMD-APP (851.4), 512MB, 479MB available)
08.02.2012 22:17:26 | | NVIDIA GPU 0 is OpenCL-capable
08.02.2012 22:17:26 | | ATI GPU 0 is OpenCL-capable
08.02.2012 22:17:26 | Einstein@Home | Found app_info.xml; using anonymous platform
08.02.2012 22:17:26 | SETI@home | Found app_info.xml; using anonymous platform
08.02.2012 22:17:26 | | Config: simulate 7 CPUs
08.02.2012 22:17:26 | | Config: use all coprocessors
08.02.2012 22:17:26 | Milkyway@Home | Config: excluded GPU. Type: ATI. App: milkyway. Device: 1

Looks like BM is recognizing it. But I cannot test it now, MW has changed his app to openCL and something is wrong with work delivery atm.

I'll check that with the albert wu's tomorrow; keep you informed.

Alexander

OK. Just remember that with more than one GPU type in the computer, that the <type/> part is required, and that the application name in the <app/> part does not include the version number and is written in all lower case. It'll just be <app>milkyway_opencl</app> or <app>milkyway_ati13</app>, or something similar.

Although the <app/> part is not required. When you do not specify which app the GPU needs to exclude, it will exclude all available applications.

---

Looks like BM is recognizing it. But I cannot test it now, MW has changed his app to openCL and something is wrong with work delivery atm.

I'll check that with the albert wu's tomorrow; keep you informed.

Alexander

Make sure you have catalyst 12.1 installed for albert, there is also poem that has an opencl app

---

CIC1=CC=C(C2=N[C@@H](CC(OC(C)(C)C)=O)C3=NN=C(C)N3C4=C2C(C)=C(C)S4)C=C1

Hi,
first part works now, MW is crunching here, but an app_info is required on both machines. Without app_info no work is distributet. People there post that gpu-detection is faulty, a ATI is detected as nVidia.

I'll check the other part tomorrow.

Good night

Update:
Basically the system is working as I want it;
nVidia crunches Einstein,
HD58xx crunches Milkyway,Albert and Seti
HD65xx (APU) crunches Albert and Seti

Share is set equal for MW, Einstein and Albert. What happens is, that MW fills the buffer (0.5 days), no work for Seti or Albert is downloaded and the second GPU, the APU, stays idle.
Need to put MW to hold, wait some time to get some Albert / Seti wu's downloaded, and set MW to work again. System needs babysitting.