Appeal No. 94-3710
Application 07/777,608
indicate that the entry in the cache contains a copy of the
data that is in the main memory.
Kaplinsky discusses presence bits in its description of
semi-associative caches, which are a compromise between a
tagged cache and an associative cache. Id., at col. 10, lines
13-14. The address for each entry in the cache contains L
label bits and I index bits. Thus, the semi-associative cache
L I
is divided into 2 sub-caches each containing 2 entries. When
checking to see if a particular word is in the cache, the L
label bits are compared against a content addressable memory.
If there is a hit, this means the sub-address space for the
particular word is mapped into a sub-cache.
The sub-cache is not filled all at once; rather, each
word is copied in when it is used (as per normal caching
principles). Thus, there is a presence bit associated with
each entry in the cache that is set only if the appropriate
word has been copied from the main memory into that entry
(i.e., the data in that entry is valid). The presence bits
are stored in a 2 x 1 RAM that is associated with each sub-I
cache. See col. 10, line 47 - col. 11, line 9.
The 2 x 1 RAM in Kaplinsky merely contains the presenceI
7
Page: Previous 1 2 3 4 5 6 7 8 9 10 11 12 Next
Last modified: November 3, 2007