Date: | Fri, 3 Jul 2009 10:23:46 +0800 |
---|---|
From: | ruzhu kao <kaoruzhu@xxxxxxxxx> |
Subject: | [Gems-users] a puzzle on the semantic of TM of ACID |
Hi all: I have read some literatures about transactional memory, which make me a little puzzle whether we should consider Consistent as TM's attribute. Especially about LogTM, it does't import C, but when I read its precursor TLTM which treats ACI as its attributes. In my opinion, consistent is an important attribute to guarantee system's correctness. My question is why recently HTMs including LogTM have not explicitly treat it as one of their attributes? Any comments will be welcome. |
[← Prev in Thread] | Current Thread | [Next in Thread→] |
---|---|---|
|
Previous by Date: | [Gems-users] Dtlb miss handling, Niran Soundararajan |
---|---|
Next by Date: | Re: [Gems-users] How can I handle non-contiguous physical address?, Jae W. Lee |
Previous by Thread: | Re: [Gems-users] Adding L3, L4 (or beyond) to the system, HyoukJoong Lee |
Next by Thread: | [Gems-users] ATMTP: Problem with strange Transaction-Aborts, Philipp Tölke |
Indexes: | [Date] [Thread] |