Carl and I have had an offline conversation to clear up some confusion. I had proposed using the 1-loop scheme with 3 or 4 way demultiplexing to facilitate pipelining. My proposal was aimed at pleasingg both the "pipeliners" and the "I have EDE hardware now" folks. However, Carl points out that although operating 3 or 4 streams in parallel achieves the pipelining, it may require more data paths than the simple 3-loop scheme. Steve
<Prev in Thread] | Current Thread | [Next in Thread> |
---|---|---|
|
Previous by Date: | Re: Attribute Ordering, Raymond Lau |
---|---|
Next by Date: | Re: Attribute Ordering, David M. Balenson |
Previous by Thread: | Re: PEM WG Meeting Minutes, Raymond Lau |
Next by Thread: | DES EDE, once more with feeling, peace |
Indexes: | [Date] [Thread] [Top] [All Lists] |