[LLVMdev] LLVM and managed languages

[Andrew Trick]

On Jul 1, 2011, at 4:41 PM, Andrew Trick wrote:
> Java bytecode bakes language semantics into opcodes as a compression technique. It's not a form that is amenable to optimization.

I mispoke slightly. "compressing" the bytecode is as much an interpreter performance optimization as it is an I/O optimization.

By "not amenable to optimization" I meant high level bytecode is not a suitable IR for an optimizing compiler.

I just assumed you're not developing an interpreted language. If so, then develop your own bytecode or use the JVM :)

-Andy