When running CPU/LLR challenges, I used to run Genefer-WR CUDA tasks alongside as they used very few CPU cycles while giving the GPU something to do. I tried the same thing when the current challenge started a couple of days ago, and to my dismay I saw that Genefer was grabbing as much CPU time as it could, competing with my LLR processes which were running the challenge. This was true for both short and WR GFN tasks.
Months ago (sorry I can't be more precise than that) GFN-WR used maybe 3% of a core at most. GFN-short used maybe 15%. Today GFN is sitting there at 50% or more. I note that the linux versions of genefer were updated in August.
Linux (ubuntu 13.10), 64 bit
3770K, HT off
I remember there's a well-discussed (on these boards) "misfeature" with Genefer OCL "hogging" a CPU core. But I repeat that I am now observing the current "hog" issue with the CUDA version, not OCL. I haven't done any OS or video driver updates in a while (maybe I should!)