You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm curious if these plugins are getting SIMD-optimized during compilation. I suspect not, as they are very CPU hungry (e.g. 2 instances of AnalogTape = 17-18% CPU). So, for instance, I might use the VA filters for a monophonic bass line but I would have to avoid too many of them at once.
I'm afraid my C++ skills are not good enough to review the plugins for this. Spacechild1 had pointed out to me on the SC forum that 8-sample loops can clue the compiler in to places where SIMD instructions could be used. This resulted in a massive speedup for me.
The text was updated successfully, but these errors were encountered:
I agree - I think this would be awesome. I don't think (this is the disadvantage of only porting other's code and not writing it yourself from the ground up) that it's not simd optimized.
If anyone has pointers on how to do it, I can look into it and mess around with it at some point. As an alternative we could also port the Surge ports of the same, they seem to have someting along those lines going on?
I'm curious if these plugins are getting SIMD-optimized during compilation. I suspect not, as they are very CPU hungry (e.g. 2 instances of AnalogTape = 17-18% CPU). So, for instance, I might use the VA filters for a monophonic bass line but I would have to avoid too many of them at once.
I'm afraid my C++ skills are not good enough to review the plugins for this. Spacechild1 had pointed out to me on the SC forum that 8-sample loops can clue the compiler in to places where SIMD instructions could be used. This resulted in a massive speedup for me.
The text was updated successfully, but these errors were encountered: