= Compiler back end
A <compiler back end> translates an <intermediate representation> into operations suitable for a target machine. Typical tasks include <instruction selection>, <register allocation>, <machine instruction>[instruction] scheduling and emission of assembly or object code. It uses facts about the target's <machine instruction>[instructions], <processor register>[registers], calling conventions and data layout.<The LLVM Target-Independent Code Generator>[Introduction and code-generation pipeline]{ref}
For example, a load from `base + 4*i` might use one indexed-addressing <machine instruction>[instruction] when the target supports that addressing mode, or require separate address arithmetic otherwise. If more values are live than <processor register>[registers] are available, <Register allocation> may spill some values to <computer memory>[memory]. Those choices change performance while having to preserve the source program's specified observable behavior. A target-independent intermediate form is therefore useful, but does not erase target-specific constraints.
Back to article page
