Clock in the center always measures dilation of rotating clocks. Rotating clock measures acceleration of clock in the center. Clocks on opposite sides of the circumference tick "at the same rate", i.e. rotating observer will see, that a clock on the opposite side of circumference neither dilates nor accelerates.
Accelerates I mean ticks faster.
Confirmed by experiments: Champeney and Moon time dilation test, Kholmetskii at all time dilation test.
It is most convenient to consider this case through Transverse Doppler Effect, when source is in the center and absorber rotates and vice versa.
Very simple animation, last episode in particular:
https://www.youtube.com/watch?v=hnphFr2Iai4
Links:
https://www.researchgate.net/publication/304781760_Specific_Features_of_Time_Dilation_During_Circular_Movement
Also the last paragraph at Mathpages
http://mathpages.com/home/kmath587/kmath587.htm
It is also interesting to consider a rotating mirror. Source in the center emits photon. Photon comes to mirror at oblique angle. Mirror gains energy of photon, i.e. photon is blueshifted at the mirror. Mirror spits photon out and loses the same amount of energy. Photon leaves mirror at oblique angle again and travels back to the source. It comes to the source at right angle and redshifts, i.e. source receives the same frequency (i.e. the same energy back) as it was released. Mirror was a "moving clock", i.e. dilates. If photon went through a center and went further, it blueshifts again. i.e. if photon was "green" (it's proper frequency) when it started from a mirror on certain point of circumference it will be "green" again on the opposite side of circumference. Sure, we consider perfect mirrors.
It is interesting to note that rotating clocks do not move relatively to each other.