Memory organisation

There are several ways to organise memories with respect to the way they are connected to the cache:

  1. one-word-wide memory organisation
  2. wide memory organisation
  3. interleaved memory organisation
  4. independent memory organisation

One-Word-Wide

The memory is one word wide and connected via a one word wide bus to the cache.

Wide

The memory is more than one word wide (usually four words wide) and connected by an equally wide bus to the low level cache (which is also wide). From the cache multiple busses of one word wide go to a MUX which selects the correct bus to connect to the high level cache.

Interleaved
Main article: Interleaved memory

There are several memory banks which are one word wide, and one word wide bus. There is some logic in the memory that selects the correct bank to use when the memory gets accessed by the cache.

Memory interleaving is a way to distribute individual addresses over memory modules. Its aim is to keep the most of modules busy as computations proceed. With memory interleaving, the low-order k bits of the memory address generally specify the module on several buses.

gollark: Mostly, but in *some* circumstances performance is more important.
gollark: Palaiologos neural networks doing all programming WHEN?
gollark: If your thing is terribly slow or your application is really widely deployed, then you can afford more programmer time.
gollark: Not *always*.
gollark: But people will be annoyed if operations are *really* slow, and for large-scale stuff you do not want to have to just buy expensive server capacity once you hit 10 requests a second.

See also

This article is issued from Wikipedia. The text is licensed under Creative Commons - Attribution - Sharealike. Additional terms may apply for the media files.