and two outputs
|
|
|
|
Note:
|
memory address (sent by CPU)
<----------------------------------->
100aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
+++**********************************
^ |<------------------------------>|
| Address for 1 memory module
|
Routing
information
|
|
Because the routing (first) bit is 1 , the entire message (except the routing bit) is forwarded onto the lower connection
|
Because the routing (first) bit is 0 , the entire message (except the routing bit) is forwarded onto the upper connection
|
Because the routing (first) bit is 0 , the entire message (except the routing bit) is forwarded onto the upper connection
|
Because the routing (first) bit is 1 , the entire message (except the routing bit) is forwarded onto the lower connection
|
Because the routing (first) bit is 0 , the entire message (except the routing bit) is forwarded onto the upper connection
|
Because the routing (first) bit is 0 , the entire message (except the routing bit) is forwarded onto the upper connection
|
|
|
|
|
|
|
|
|
|
|
|
This kind of conflict happens inside the switching fabric and it is called internal blocking requests
|
|
If the routing (first) bit is equal to 0 (zero), then the message is routed towards the TWO UPPER Banyan switches in the SECOND stage
If the routing (first) bit is equal to 1 (one), then the message is routed towards the TWO LOWER Banyan switches in the SECOND stage
|
|
If the routing bit (second bit in address) is equal to 0 (zero), then the message is routed towards the TWO UPPER Banyan switches in the THIRD stage
If the routing bit (second bit in address) is equal to 1 (one), then the message is routed towards the TWO LOWER Banyan switches in the THIRD stage
|
(It's so elementary, my dear Watson :-))
Here is how you connect a 16x16 Delta MIN:
|
n log(n)
----------
2
|