Introduction

In this segment of the course we consider compilation issues for loops involving arrays:

$\circ$

How execution order of a loop is constrained,

$\circ$

How a compiler can extract dependence information, and

$\circ$

How this can be used to optimise a program.

Understanding and transforming exec$^{\rm n}$ order can exploit architectural features:

• Pipelined, superscalar and VLIW processors
• Systems which rely heavily on caches.
• CPUs with special instructions for vectors.
• Multiprocessors.