After some spinning (lost some WU), i got CUDA AND Astropulse running on my Laptop. Its a DELL E6500 running on NVIDIA Quadro NVS 160M.

1. updated to NVIDIA driver 178.08 (beta) with modified nv_cuda.inf.

-> CUDA WU's came down, but olny CUDA.

2. updated to Boinc 6.6.4 and modifed cc_config.xml to <ncpus>2</ncpus> !!!

(funny: if i set that to 3, i get 3 CPU WU and one GPU WU running at the same time: 1st WU using 50% CPU and 2nd and 3rd WU each on around 25% CPU and one WU on 100% GPU)

-> but here we go: two AP WU on CPU (dual core) and one CUDA WU on GPU

But now i am completly on SETI stock apps...
When i look at the processor time in the boinc client, the CUDA WU counts up to only 2 or 3 minutes, but it takes more than 2 hours to get there ???

Any chance to get some optimized stuff running?

I tested some mod stuff but this seems to be dependent on newer NVidia drivers (1.8.0) which are not yet av for this type of Laptop.

Question: Any chance to get some optimized apps to work? Maybe opt AP and stock CUDA?
Thanks in advance...
TOM

---

2. updated to Boinc 6.6.4 and modifed cc_config.xml to <ncpus>2</ncpus> !!!

For versions above 6.5.0 you do not need the ncpus option in cc_config.xml anywmore. BOINC will do this automatically.

(funny: if i set that to 3, i get 3 CPU WU and one GPU WU running at the same time: 1st WU using 50% CPU and 2nd and 3rd WU each on around 25% CPU and one WU on 100% GPU)

Yes, it's a DEBUG option to temporarily run work on more CPUs than you actually have. You should not use it to set the amount of CPUs you have.

But now i am completly on SETI stock apps...
When i look at the processor time in the boinc client, the CUDA WU counts up to only 2 or 3 minutes, but it takes more than 2 hours to get there ???

BOINC still only counts the CPU time, it will not show the GPU time. The tasks run on your GPU, not on your CPU. The time increasing for the CPU is only showing the work the CPU does, which is the transferring of the task from the disk to the video RAM and back. The GPU completely calculates the task.

I tested some mod stuff but this seems to be dependent on newer NVidia drivers (1.8.0) which are not yet av for this type of Laptop.

See this FAQ for alternative drivers. You can find it in the requirements section of the FAQ.

---

Addendum: since yesterday there are also CUDA Enhanced 6.08 WU, Enhaced 6.03 WU and also AP 5.0 WU and AP 5.03 WU on my Laptop.
Any idea of how an app_info.xml has to look like to handle that with optimized apps ???
TOM

---

Check the thread: AK V8 + CUDA MB team work mod

---

I guess i can't use this mod because my nVidia drivers are on 178.08
I try do make my own homebrew app_inf.xml:
It will use CPU AK_v8_win_SSE4.1.exe on 5.28 and 6.03 WU
It will use CPU ap_5.00r69_SSE3.exe on AP 5.00 WU
It will use CPU astropulse_5.03_windows_intelx86.exe on AP 5.03 WU
It will use GPU setiathome_6.08_windows_intelx86__cuda.exe on 6.08 WU

I'll try to use an app_info.xml like this:

<app_info>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>AK_v8_win_SSE41.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>528</version_num>
<file_ref>
<file_name>AK_v8_win_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<file_ref>
<file_name>AK_v8_win_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>astropulse</name>
</app>
<file_info>
<name>ap_5.00r69_SSE3.exe</name>
<executable/>
</file_info>
<file_info>
<name>libfftw3f-3-1-1a_upx.dll</name>
<executable/>
</file_info>
<app_version>
<app_name>astropulse</app_name>
<version_num>500</version_num>
<file_ref>
<file_name>ap_5.00r69_SSE3.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>libfftw3f-3-1-1a_upx.dll</file_name>
</file_ref>
</app_version>
</app_info>


<app_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<file_ref>
<file_name>setiathome_6.08_windows_intelx86__cuda.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft.dll</file_name>
</file_ref>
<file_ref>
<file_name>libfftw3f-3-1-1a_upx.dll</file_name>
</file_ref>
</app_version>
</app_info>

<app_info>
<app_version>
<app_name>astropulse</app_name>
<version_num>503</version_num>
<file_ref>
<file_name>astropulse_5.03_windows_intelx86.exe</file_name>
<main_program/>
</file_ref>
<file_info>
<name>libfftw3f-3-1-1a_upx.dll</name>
<executable/>
</file_info>
</app_version>
</app_info>

correct?

Tom

---

I guess i can't use this mod because my nVidia drivers are on 178.08
I try do make my own homebrew app_inf.xml:
It will use CPU AK_v8_win_SSE4.1.exe on 5.28 and 6.03 WU
It will use CPU ap_5.00r69_SSE3.exe on AP 5.00 WU
It will use CPU astropulse_5.03_windows_intelx86.exe on AP 5.03 WU
It will use GPU setiathome_6.08_windows_intelx86__cuda.exe on 6.08 WU

I'll try to use an app_info.xml like this:

<app_info>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>AK_v8_win_SSE41.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>528</version_num>
<file_ref>
<file_name>AK_v8_win_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<file_ref>
<file_name>AK_v8_win_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>astropulse</name>
</app>
<file_info>
<name>ap_5.00r69_SSE3.exe</name>
<executable/>
</file_info>
<file_info>
<name>libfftw3f-3-1-1a_upx.dll</name>
<executable/>
</file_info>
<app_version>
<app_name>astropulse</app_name>
<version_num>500</version_num>
<file_ref>
<file_name>ap_5.00r69_SSE3.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>libfftw3f-3-1-1a_upx.dll</file_name>
</file_ref>
</app_version>
</app_info>


<app_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<file_ref>
<file_name>setiathome_6.08_windows_intelx86__cuda.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft.dll</file_name>
</file_ref>
<file_ref>
<file_name>libfftw3f-3-1-1a_upx.dll</file_name>
</file_ref>
</app_version>
</app_info>

<app_info>
<app_version>
<app_name>astropulse</app_name>
<version_num>503</version_num>
<file_ref>
<file_name>astropulse_5.03_windows_intelx86.exe</file_name>
<main_program/>
</file_ref>
<file_info>
<name>libfftw3f-3-1-1a_upx.dll</name>
<executable/>
</file_info>
</app_version>
</app_info>

correct?

Tom

You can have only one pair of <app_info> ... </app_info> tags in an app_info.xml file. So this file will NOT work. Within the file, the app_name for the AP500 and AP503 should be different. The reason CPU and GPU cannot work together at present with an app_info is that the app_name for both is the same - "setiathome_enhanced".

F.

---