4 ms·
it would depend on whether the frontend generates optimal LLVM IR for the given code. For example, if a language implemented multiplication through repeated add
by wscp-dev 3y ago
it would depend on whether the frontend generates optimal LLVM IR for the given code. For example, if a language implemented multiplication through repeated addition (though I think LLVM may be able to optimise that itself), the result would be suboptimal to generating a single IR code for multiplication.