Mine is driver version 10.18.10.3621, so very close. But maybe I'll look for a newer one in the morning.

Launching more work now for all 3 gpu types on all 3 OS systems.

Thanks,
-Uplinger

This beta will have a total of 34,491. This will be sent out in pairs to start, but once machines start validating on it, then it'll work similar to other OPN1 results.

Thanks,
-Uplinger

Thanks for providing that information!

The intel GPU WU on my macbook finished successfully within ~1h. ETA timing was on spot. Now in PV.

For this beta run on my NVIDIA cards, I tried an app_config file settings to let every GPU crunch 2 tasks concurrently. GPU util while still having short drops to ~0%, is much less 'spiky' and the load is sustained longer on the ~100% level. I expect throughput to be higher with this setting compared to only allowing one WU at a time.

Edit: I tried running 3 at a time now, and everthing is running smoothly. Additionally, I tried suspending the tasks during computation and then conitinuing them, not always being sure if the tasks were conitinued on the same GPU or handed over.... But everthing seems fine and is working as expected

Is there a way to run 2 WUs on Nvidia and only 1 on Intel? Have tried different configs but can´t get it to work, either both cards run 1 or 2 WUs.

My current app_config

<app_config>
<project_max_concurrent>10</project_max_concurrent>
<app>
<name>beta29</name>
<plan_class>opencl_nvidia_102</plan_class>
<gpu_versions>
<gpu_usage>.5</gpu_usage>
<cpu_usage>.9</cpu_usage>
</gpu_versions>
</app>
<app>
<name>beta29</name>
<plan_class>opencl_intel_gpu_102</plan_class>
<gpu_versions>
<gpu_usage>1</gpu_usage>
<cpu_usage>.9</cpu_usage>
</gpu_versions>
</app>
</app_config>