Code Generation for Arrays and Records
What you'll learn: How compilers translate array and struct field accesses into memory address calculations using base-plus-offset arithmetic.
When you write arr[5] or person.age in high-level code, the compiler must translate these into actual memory addresses. Unlike simple variables that live at a fixed location, arrays and records (structs) require computed addresses based on the element index or field being accessed.
The Base-Plus-Offset Model
Think of memory like a street of numbered houses. An array's base address is the street corner, and each element sits at a predictable offset from that corner.
For arrays, the formula is:
address = base + (index × element_size)
If arr starts at address 1000 and holds 4-byte integers, arr[5] lives at 1000 + (5 × 4) = 1020.
For records/structs, each field has a fixed offset from the structure's start:
address = base + field_offset
If a Person struct starts at 2000, and the age field is 8 bytes into the structure, person.age is at 2000 + 8 = 2008.
Code Generation Pattern
The compiler generates three-address code like:
t1 = index × 4 // multiply by element size
t2 = base + t1 // add to base address
result = *t2 // dereference to load value
For nested structures or multidimensional arrays, these calculations stack: each level adds another offset computation.
Key Takeaway: Code generation for arrays and records transforms symbolic accesses into explicit address arithmetic using base pointers and compile-time or runtime-computed offsets—the foundation of all structured data access in compiled code.