RE vs VST efficieny (platform specific optimizations)?

[WillyOD]

So many people still posting about VST vs REs performances. They don't seem at all directly comparable to me, but it got me wondering...

How much of the efficieny do REs lose against VSTs because they're multiplatform? And how much because of their Reason specific complexity?

How limiting is the SDK? Could you for example implement platform specific inline Assembly optimizations for REs while still providing a more basic pseudo-fallback-routines for other platforms? Sorry if these are silly questions, but since I don't have access to the SDK I have no idea about the RE interface/implementation or its limitations.

[WillyOD]

I guess easier way of putting it; Is it possible to provide OS-platform specific optimizations inside REs?

[eXode]

VSTs can be better optimized than REs as far as I understand.

This has to do with the fact that the RE code is platform agnostic, and that RE's (currently) are sandboxed. With VST the developer can do optimizations on hardware level, etc. Assembly optimizations is not possible in REs (I believe it was one points that Urs of u-he brought up). This is put simply, and as far as I understand it.

[etyrnal]

So many people still posting about VST vs REs performances. They don't seem at all directly comparable to me, but it got me wondering...

How much of the efficieny do REs lose against VSTs because they're multiplatform? And how much because of their Reason specific complexity?

How limiting is the SDK? Could you for example implement platform specific inline Assembly optimizations for REs while still providing a more basic pseudo-fallback-routines for other platforms? Sorry if these are silly questions, but since I don't have access to the SDK I have no idea about the RE interface/implementation or its limitations.

and here comes the nightmare that is vst... fragmentation, etc...

Sent from my XT1575 using Tapatalk