Hehe, didn't want to annoy you.Please note that I am full-time developer with an university degree diploma. I am programming audio stuff since 27 years now and developed over 20 audio-products. I really should know how things should work
My guess is FLStudio does variable size processing to make vst parameter automation sample-exact in vst2 plugins, so...2) When a block is processed there is a context-switch. A context switch always has lots of overhead. This is unavoidable. When a block consists of just 1 sample size (yes FL Studio does this!) lots of CPU is wasted and the cache is also messed up. This results in CPU spikes and possible dropouts.
https://en.wikipedia.org/wiki/Context_switch
