Kayvon mentioned that the advantage of implementing these abstractions at the compiler level instead of writing, say, a C++ library is that it can make some optimizations. What are some examples of such compiler-identified optimizations that wouldn't be possible with a library?
Kayvon mentioned that the advantage of implementing these abstractions at the compiler level instead of writing, say, a C++ library is that it can make some optimizations. What are some examples of such compiler-identified optimizations that wouldn't be possible with a library?
edit: nvm, discussed later in the lecture