Date: | Tue, 18 Dec 2007 21:31:59 -0600 |
---|---|
From: | "Mike Marty" <mike.marty@xxxxxxxxx> |
Subject: | Re: [Gems-users] MOSI_SMP_bcast |
Hi, Nope. Physical time is meaningless. As long as messages from the root are ordered, we get sequential consistency in this protocol. It is the appearance that matters. A later store cannot become visible to some other processor because of the ordering at the root. Even in bus-based systems, a processor can buffer snoops and different processors may not actually carry out snoops in the same physical time. As long as a processor processes all snoops before initiating a bus transaction, the appearance of SC is maintained. These issues are covered in detail in Scheurik's PhD thesis.
Ordering-Point can actually work fine without the explicit Ack on a single ring, or multiple interleaved rings that are synchronous. |
[← Prev in Thread] | Current Thread | [Next in Thread→] |
---|---|---|
|
Previous by Date: | Re: [Gems-users] How can I run multi-programmed benchmarks?, 张量 |
---|---|
Next by Date: | [Gems-users] Debug Information and Sequencer, Dimitris Tsaliagos |
Previous by Thread: | [Gems-users] MOSI_SMP_bcast, Niket |
Next by Thread: | [Gems-users] MOSI_SMP_bcast finite state machine, nikiforg |
Indexes: | [Date] [Thread] |