Most of the ctimer configuration is taken up with debugging and profiling functionality. Whilst real-time applications could use the clock counter for timing, they aren't really any use for computational blocks once the debugging and profiling is done with.
But with a small change they could be: add a timing mode which counts how many times the timer value register is read.
This could be used to implement up to 2x atomic counters per core which is a useful feature for implementing various parallel algorithms efficiently.