| Date: | Tue, 02 Oct 2007 22:10:16 -0700 |
|---|---|
| From: | cge@xxxxxxxx |
| Subject: | [Gems-users] MOESI_SMP_LOGTM protocol question |
Hi,in the MOESI SMP LOGTM directory cache protocol(version 1.3), there is a following transition:
transition(MM, Fwd_GETS_T, O) {
e_sendData;
l_popForwardQueue;
}
Since Fwd_GETS_X event will be generated instead of Fwd_GETS_T if the
cache entry's
m_write field is false, the above transition seems to imply that a
cache entry can be in
a MM state with the m_write field to be false.Is a cache entry's m_write always true if the cache entry is in MM state because MM means the entry is dirty and locally modified? Thanks, Ken |
| [← Prev in Thread] | Current Thread | [Next in Thread→] |
|---|---|---|
| ||
| Previous by Date: | Re: [Gems-users] Tourmaline and scheduling, Matthew James Horsnell |
|---|---|
| Next by Date: | Re: [Gems-users] MOESI_SMP_LOGTM protocol question, Kevin Moore |
| Previous by Thread: | Re: [Gems-users] MI_Example & MOSI_SMP_bcast protocols, Mike Marty |
| Next by Thread: | Re: [Gems-users] MOESI_SMP_LOGTM protocol question, Kevin Moore |
| Indexes: | [Date] [Thread] |