Hasn’t obviously failed yet in a fair bit of testing, but will continue to test. The CPU spikes are much more of a problem for me it seems
The drift correction is definitely causing glitches for me. Could be a buggy algorithm? …or maybe it’s just sporadically CPU intensive.
Only other solution would be to ditch the USB audio devices’ USB audio connection, which wouldn’t be ideal.
If a device is taking its clock triggers from the Mac’s USB and synching to that (which some do seem to do), then things should be OK unless those clock triggers get missed. Bearing in mind that many devices will synch to the host clock anyhow, if they’re providing analogue outputs.