Week 5

1. Which are the two forms of locality?
   • Spatial and temporal
2. What is meant by a cache block?
   • The bytes read from memory on a cache miss are collectively called a cache block
3. A cache is divided into a number of sets, and when an address is searched all rows in a particular set are checked to see if that row con- tains the searched address and needed data. What is a cache called if it is described as follows:
   • The cache has only one set - fully associative
   • Each set has four rows - set-associative cache
   • The number of sets and rows are equal - direct-mapped cache
4. When there are multiple rows in a set, are they searched in sequence or are the hardware address comparators duplicated so that they work in parallel?
   • The hardware address comparators are duplicated so that they work in parallel
5. What is the cache block size on the power machine?
   • 128 bytes
6. Modern optimizing compilers use a representation which makes it relatively easy analyze which statement has modified a variable. What is this representation called?
   • SSA (Static Single Assignment) form
7. What is the purpose of instruction scheduling?
   • To improve performance by reducing the number of pipeline stalls suffered during execution