Thread-Level Parallelism - Computer Architecture: A Quantitative Approach - page 420

shown have invalid blocks. For each of the sequences below, identify which nodes (chip/
processor) receive each request and invalidate.
a. [10] <5.4> P0,0: write 100 <-- 80
b. [10] <5.4> P0,0: write 108 <-- 88
c. [10] <5.4> P0,0: write 118 <-- 90
d. [10] <5.4> P1,0: write 128 <-- 98
5.11 [25] <5.4> Exercise 5.3 asked you to add the Owned state to the simple MSI snooping
protocol. Repeat the question, but with the simple directory protocol above.
5.12 [25] <5.4> Discuss why adding an Exclusive state is much more difficult to do with the
simple directory protocol than it is in a snooping protocol. Give an example of the kinds of
issues that arise.
Case Study 3: Advanced Directory Protocol
Concepts illustrated by this case study
■ Directory Coherence Protocol Implementation
■ Coherence Protocol Performance
■ Coherence Protocol Optimizations
The directory coherence protocol in Case Study 2 describes directory coherence at an ab-
stract level but assumes atomic transitions much like the simple snooping system. High-per-
formance directory systems use pipelined, switched interconnects that greatly improve band-
width but also introduce transient states and nonatomic transactions. Directory cache coheren-
ce protocols are more scalable than snooping cache coherence protocols for two reasons. First,
snooping cache coherence protocols broadcast requests to all nodes, limiting their scalability.
Directory protocols use a level of indirection—a message to the directory—to ensure that re-
quests are only sent to the nodes that have copies of a block. Second, the address network of a
snooping system must deliver requests in a total order, while directory protocols can relax this
constraint. Some directory protocols assume no network ordering, which is beneficial since it
allows adaptive routing techniques to improve network bandwidth. Other protocols rely on
point-to-point order (i.e., messages from node P0 to node P1 will arrive in order). Even with
this ordering constraint, directory protocols usually have more transient states than snooping
protocols. Figure 5.39 presents the cache controller state transitions for a simplified directory
protocol that relies on point-to-point network ordering. Figure 5.40 presents the directory con-
troller's state transitions.