I guess I'm old, the macOS behaviour is more in line with my expectations.
But this got me thinking, how feasible it would be to tie the various clock systems of a computer to some reference clock, like 10 MHz GPSDO? Obviously it wouldn't improve the granularity, but you could ensure that the timestamps are actually accurate. Because otherwise I doubt that random computer clock would be accurate down to 32ns even with NTP.
Getting a 10MHz PPS signal requires specialized expensive hardware and typically doesn't scale well to cover every server. That sort of thing is best left to extreme applications with FPGAs or ASICs. In particular the 10MHz version; a lot of commodity hardware only supports the 1Hz one.
A major problem in synchronization of the system clock is PCIe. Hardware can timestamp PPS signal or PTP/NTP packets with accuracy of a few nanoseconds if everything is well compensated, but the PCIe bus between the CPU and the timestamping HW has a latency of hundreds of nanoseconds with potentially large asymmetry, degrading the accuracy significantly.
Measuring that error is difficult. A possibility is to run a program periodically making random reads from memory (avoiding the CPU cache) to generate a PPS signal on the memory bus, which can be observed with a scope. There is a lot of noise due to other memory activity, RAM refresh, etc. From the configured memory speed and tRCD+tCL timings the uncertainty of the error can be reduced.
This might improve with new hardware. There is a feature called Precision Time Measurement (PTM), which is a hardware implementation of an NTP-like protocol in PCIe, but so far I have seen this working only on some onboard NICs.
But this got me thinking, how feasible it would be to tie the various clock systems of a computer to some reference clock, like 10 MHz GPSDO? Obviously it wouldn't improve the granularity, but you could ensure that the timestamps are actually accurate. Because otherwise I doubt that random computer clock would be accurate down to 32ns even with NTP.