discussion forum
message
| Name: |
Oliver |
| eMail: |
Follow.Me@gmx.net |
| Date: |
June 17, 2003 at 07:22:56 |
| Subject: |
Re: Hyperthreading and 'lock' prefix |
| In Reply To: |
Re: Hyperthreading and 'lock' prefix by Stephan Keil on June 17, 2003 at 06:22:33 |
| Text: |
> He also explains, that LOCK only locks the cache, ... Yes, since the PPro. > ... if the affected value is locally cached. So for performance > reasons it might be useful to ensure that the value is cached. I hardly doubt that any x86 SMP-system would do that modification on another CPU's cache. I belive it's the same like with the unLOCKed ver- sions of the uniprocessor-atomic operations: when the value is in another CPU's cache, it's transferred to the local cache and invalidated in the remote cache; and if there's a redundant copy in any remote cache, it's simply invalidated - just the MESIsh way. |
post a followup message
(Be nice... or be blocked. Be technical... or be erased.)
currently posted followup messages
(You may have to press the RELOAD button of your browser.)
|