In the late 1970's there was published high temp fatigue curves for P22 , 314 H, 316 H, 347H, associated with researchon the Clinch River Breeder reactor.  Look in the ASME journals from that period, but I suspect the data is probably for parent metal and not weld metal.

The properties of the weld metal are associated with the specific weld procedure for each tested sample, such as preheat, interpass temp, and PWHT procedures, so it would be very difficult to demonstrate high repeatability unless the welding process is totally mechanized.

There are also newer results published in the online tech  journal OMMI at <www.ommi.co.uk> for P22 and P91

For the P22 fatigue data, see Brinkman et al:"Elevated temperature fatigue behavior of 2 1/4Cr 1 Mo steel", J Press vess tech, trans ASME Nov 1975 pp252-257. This formed the basis of the fatigue curves in the ASME code case N-47.

There are newer fatigue curves in the EU PED, such as in EN 12952-3:2001 annex B. The fatigue curves are meant to be applied to the parent material , while the weld zone may be treated completely differently ( as in BS 5500), where the failure of a weldment uses fracture mechanics at a pre-existing microcrack.