Scalar Architecture:  Cache Organization

• The goal of scalar optimization:

  • Spatial locality of reference (use all words in cache line)
  • Temporal locality of reference (use same cache line)

Words in Memory

Example Cache L2 on O2K

(e.g. 8 MB

or 2097152 words)

cache lines

in memory (32 words)

Load instruction

(ld)

for 1 word

cache line

transfer

• Cache hit will load word from cache

• Cache miss will load cache line from memory

Previous slide

Next slide

Back to first slide

View graphic version