I have 3 NV GPU's running beta test MB V8 with 4000+ WU's returned W/O errors ( as far as I can tell)

Is there a app_info.xml available so I can run the GPU's at more than 25%

Ed F

---

You could probably run a simple app_config.xml to get more than 1 work unit per card.

of course you will need to create app versions for each of the plan classes, cuda 32, cuda 42, cuda 50 and opencl_nvidia_sah

Simple one would look like

<app_config>
<app_version>
<app_name>setiathome_v8</app_name>
<plan_class>cuda50</plan_class>
<avg_ncpus>0.4</avg_ncpus>
<ngpus>0.5</ngpus>
</app_version>
<app_version>
<app_name>setiathome_v8</app_name>
<plan_class>cuda42</plan_class>
<avg_ncpus>0.4</avg_ncpus>
<ngpus>0.5</ngpus>
</app_version>
<app_version>
<app_name>setiathome_v8</app_name>
<plan_class>opencl_nvidia_sah</plan_class>
<avg_ncpus>0.4</avg_ncpus>
<ngpus>0.5</ngpus>
</app_version>
</app_config>

This of course is a rough draft. Can you do an app_info, yes but you will need to specify all the different sections that the app needs to run and where they are. You know the all the libfftw3f, multibeam_Kernels_r3330, etc. If you miss any of those or use the wrong one, it probably won't work or will throw errors.

If you just want to increase the number, this is how I would go.

Have I built an app_info,yes but it took me a while to figure out the correct sequence and I have to monitor my results to make sure I don't get any errors.

This way you are only increasing the number of work units. Keep an eye on your temps.