WRITE MISS POLICY:

•   Write allocate
•   fetch on write
•   fetch entire block, then write word into block
•   allocate a new block on each writ
•   allocate block, but don't fetch
•   no-fetch
•   requires valid bits per word
•   Write no-allocate
•   more complex eviction
•   don't allocate a block if it is not already in the cache
•   naturally used by write through since we need update main memory anyway
•   Write invalidate
•   write around the cache
•   instead of update for write-through