I had re-read the whole thread, and what you are describing as an issue, is the supposed unbalanced use of the 2 cores of your CPU. From all test I have performed, I can't replicate that, except maybe for the second image of post #7: http://www.pclinuxos.com/forum/index.php/topic,94311.msg794010.html#msg794010
As you may have seen things appear to be different using different kernel, different governors, and very probably different running applications. BFS scheduler coupled to "ondemand governor" appear to switch between the 2 core very often (say 1 or more time per second), while the same using "powersave governor" switch approx every 15 secs. CPU Load it's however balanced, in different way but still balanced.
The graph from my tests using gkrellm, are as large as 4 minutes, the large of your ksysguard test, as per default in FM, is large 1 minutes, it's unlikely you can see a switch between the 2 core if it happen say after 3 minutes.
Even if a process never switch to the second core, what's the problem ? The other core remain available for use from other tasks, moreover if instead of your test (while true; do true; done) you run a multithreaded applications like Chromium or Firefox to name a couple, you will find both cores working simultaneously.
Going to test different systems like LXDE vs KDE, is likely to introduce even more variables into the "issue", this is not the right path IMHO to find a root cause of any problem.
So, if it's me I would go with the latest kernel only (either .bfs or without the CK patch, depending on your requirements), and if a specific and a real issue arise, report back about.
From another point of view, kernel is a piece of software used across all distributions, and if an issue exist it will be found from one of the many "kernel developers", surely they have better tools and knowledge to test a possible issue like the one discussed here.