Shortly adding to the observations in earlier threads about the wonky GPU util due to higher idle teams. So far I have run 3 WUs concurrently on each of my GPUs, but with the new batches 3 WUs resulted in a much lower and less stable load. To combat that I adjusted that config upwards to allow 4 WUs on each one but that is not doing much for me. Before the current stress test WUs, 3 WUs usually loaded my GPUs very consistently at 95-100%, now I am seeing on average 60-65% as measured by GPU-Z on Win 10 over the last hour. (and that is with an increased load via the 1 additional WUs run concurrently, and a higher need for CPU support)

thanks uplinger, any comment about what's going on with the low GPU utilization (lots of GPU idle time) of the 5-digit batches? you had mentioned that you though they should run fast.

I even confirmed that the process is constantly comming on and off the GPU. you can catch times running nvidia-smi where it shows the wcg application isnt even running on the GPU, while BOINC shows it running. and it'll constantly pop in and out. this is much different than all the tasks before, where even if the sub jobs were starting and stopping, nvidia-smi still recognized that the application was running on the GPU.

Yup, validates fine. Also I can mention that all my stone age GPU's crunch these larger WU's just fine, and they validate OK.
(GTX980 Strix, iGPU HD4600, and GTX660M)

Starting to see a few 5 digit batch resends. Kind of a smorgasbord of errors and some user aborted but 0 invalids.YAY!

https://www.worldcommunitygrid.org/forums/wcg/changesignature