Hardware vendors often assume that system integration revolves around their own devices. This clock synchronizes devices despite such assumptions.
The clock is orders of magnitude more accurate than conventional methods in Microsoft Windows, even without real-time priority or busy-waiting. A Linux master clock provides a stable NTP time base. Slave clocks, Linux or Windows, synchronize to the master clock by several mechanisms. Measuring round-trip ping times corrects for network latency. In the slave's high-resolution clock, drift is predictively compensated for by second-order curve fitting while wraparound and jitter (e.g., from PCI bus contention) is trapped. Performance degrades gracefully and measurably on unreliable networks. Several phase-locked loops, within each slave and between slave and master, guarantee performance.