AMD athlon 64 User Manual

Page of 48
Appendix A
39
Performance Guidelines for AMD Athlon™ 64 and AMD Opteron™
ccNUMA Multiprocessor Systems
40555
Rev. 3.00
June 2006
Appendix A
The following sections provide additional explanatory information on topics discussed in the previous 
sections of this document.
A.1
Description of the Buffer Queues
Figure 16 shows the internal resources in each Quartet node. The memory controller (MCT), the 
System Request Interface (SRI) and the Crossbar (XBar) on each node all have internal buffers that 
are used to queue transmitted transaction packets. Physically each of these units only has two kinds of 
buffers, command and data. 
Figure 16.
Internal Resources Associated with a Quartet Node
Consider the buffers that come into play in the interface between the XBar and the HyperTransport™ 
links.
Each node has an incoming and outgoing coherent HyperTransport link on its XBar to every other 
node but one in the system. Node 0 has an outgoing link that allows it to send data from node 0 to 
node 1. Likewise node 0 has an incoming link from node 1 to node 0. The two links together can be 
considered as one bidirectional link.
Each node has HyperTransport buffers in its XBar that are used to queue up the packets that are going 
to be sent on the outgoing link. The sending node does not send the packets until the receiving node is 
ready to receive them.
Now Consider the buffers that come into play in the interface between the XBar and the MCT. 
Packets to be transmitted from the XBar to the MCT are queued in the ‘XBar-to-MCT” buffers. 
HT = HyperTransport™
cHT = coherent HyperTransport
C0
C1