Register to Register Architecture

• All vector operations occur between vector registers
• If necessary, operands are fetched from main memory into a set of vector registers (load-store unit)
• SIMD based on this
•