No performance improvement with FloatVectorOperations

I use Visual Studio 2022 also by Microsoft, duh as I am sure you’re aware off. What compiler do you use, that makes the “-O3” option available in ProJucer?

That’s the Mac Xcode flag. I think gcc uses the same flag.

Microsoft also offers the clang, but I think it is not selected by default in VisualStudio, but you can select it.

MSVC doesn’t have the O3 flag.
The flag you’re looking for is either /O2 or /O2 /fp:fast.

I’m not sure why the Projucer uses /Ox as the default - IMO this is something that needs to be changed by the team.

That flag will produce slower results than the previous flags I’ve mentioned.

Also I would recommend using Clang on Windows to squeeze even more performance cycles out of your compiler.