Its too early to tell for sure what the impact of this might be. It’s certainly not good and there will be some performance hit.
As far as Cantabile is concerned I don’t think it will be too bad. From what I can understand the impact will be on kernel mode calls and Cantabile’s audio engine is already very lean in this respect - it makes practically no kernel mode calls from the audio thread.
The bigger question might be the impact on plugins - depending how well the plugins are engineered and what synchronization and locking mechanisms they use the impact could be considerable. I would expect though most quality plugins do that same as Cantabile and avoid blocking kernel mode calls as much as possible.
Still on the audio engine other probably impacted areas:
- Waking the audio thread on interrupt from audio driver. I suspect this will certainly be impacted but I wouldn’t imagine the impact to be huge - this is outside Cantabiles control and comes down to the driver mostly.
- ISR and DPC - these will almost certainly be impacted and it might require disabling more devices while performing live.
Outside the audio engine I wouldn’t be too concerned - unless your machine is already running at 100% CPU all the time I doubt you’ll notice much difference.
I mentioned Cantabile makes practically no kernel mode API calls - the main exception is when running in multi-core mode where:
- it uses a critical section lock to synchronize access to a work item queue - but it has spin count on it and in theory shouldn’t ever hit the kernel.
- it signals one event at the start of the audio cycle to wake the worker threads.
TL;DR: it’s certainly not good but I’m only slightly concerned. It might mean increasing buffer sizes slightly and perhaps some other tweaks. It might also highlight some badly written plugins. Fingers crossed - I guess we’ll find out soon enough.
Disclaimer: everything above is one big guess.