Open nine7nine opened 3 months ago
hmmmm.... this may not be Wine-NSPA specific: I rolled back quite far (before some significant changesets), along testing with/without librtpi (as well) and dxvk...
Now I am wondering if it could be kernel related? Possibly the EEVDF / DL servers stuff -- that I need to update on the next release... In any case: it seems highly unlikely this is a Wine problem. Regardless, still investigating -- and we'll see if it still happens when I update the kernel (just waiting on a new RT patch + the EEVDF / DL servers patchwork).
It'll get resolved -- but at least it doesn't look caused by Large/Huge Pages support, librtpi, iosb patchwork, RT changes, or any of the locking and thread safety stuff I have been working on. phew ;-)
https://www.phoronix.com/news/Linux-Completing-EEVDF
Looks like the EEVDF and DL Servers kernel code is about to be dropped into mainline.
So I should be able to give that a test and see if it fixes things up with Kontakt - if not, more digging. sigh.
EDIT: The above patchwork has been merged for 6.11-rc2 - So when that is released, I can test again.
The gist:
I stopped using Kontakt 6 a while back, but recently I was using it and noticed a regression where it periodically causes a cpu spike (at some kind of interval). I tried reverted ome obvious candidates that may have introduced thi issue -- but so far; no dice... At first, I thought it was Ableton Live specific - but it's not...
Soooo...
I will need to do a bisect and revert to a version where it doesn't produce this spike, then work through commits && builds until I see it introduced (fun!). I haven't noticed any other VSTs affected by this, it really seems to be a Kontakt-specfic issue.
Anyway, in the coming days I will try to dedicate some time into tracking down the problematic change.