| Date: | Mon, 22 Jan 2007 13:51:15 +0800 |
|---|---|
| From: | "hongxia sun" <sunhx.seraph@xxxxxxxxx> |
| Subject: | Re: [Gems-users] GEMS state transition problem |
|
"L2_MigrationDone" is "L2_SomeEvent" or "L2_SomOtherEvent"?
When does the "L2_SomeEvent" and "L2_SomeOtherEvent" happen?
Please pay attention that the transition "1635 1 0 L2Cache L2_SomeEvent L2_NP>L2_IM [0x400, line 0x400]" happened at time 1635.
While the invalid transition also try to happen at time 1635 "Warning: in fn TransitionResult L2Cache_Controller::doTransitionWorker(L2Cache_Event, L2Cache_State, L2Cache_State&, const Ad
dress&) in generated/MSI_MOSI_CMP_directory/L2Cache_Transitions.C:1652: g_eventQueue_ptr->getTime() is 1635". At cycle 1635, the cache line state keeps as L2_NP but not L2_IM, therefore, event "L2_MigrationDone" cause an invalid transition.
在07-1-22,Lei Yang <lya755@xxxxxxxxxxxxxxxxxxxx> 写道:
|
| [← Prev in Thread] | Current Thread | [Next in Thread→] |
|---|---|---|
| ||
| Previous by Date: | [Gems-users] GEMS state transition problem, Lei Yang |
|---|---|
| Next by Date: | [Gems-users] A tricky design of delay restarting an aborted transaction in LogTM, 郭锐 |
| Previous by Thread: | [Gems-users] GEMS state transition problem, Lei Yang |
| Next by Thread: | Re: [Gems-users] GEMS state transition problem, Lei Yang |
| Indexes: | [Date] [Thread] |