discussion forum
message
| Name: |
Stephan Keil |
| eMail: |
Stephan.Keil@gmx.de |
| Date: |
June 10, 2003 at 09:09:36 |
| Subject: |
Re: Hyperthreading and 'lock' prefix |
| In Reply To: |
Re: Hyperthreading and 'lock' prefix by james on June 10, 2003 at 04:01:12 |
| Text: |
|| the x86 instruction set supports the 'lock' prefix for various || instructions (e.g. inc, add) to guarantee atomicity in multiprocessor || environments, i.e. multiple _physical_ processors. But is this || prefix also necessary for hyperthreading processors, i.e. two || logical processors within one physical processor? | | Yes. | | I mean, how else is virtual processor #2 gonna know | that it shouldn't write to a location that virtual | processor #1 is currently operating on with a read- | modify-write operation?Both processors share the same pipeline. If the pipeline would care about read-modify-write conflicts regardless of the executing thread (as it does for a single thread), then it would be safe to omit the 'lock'. So my question refers more to the current implementations of hyperthreaded processors (P4-XEON?!): does the P4 pipeline care about those conflicts regardless of the executing thread? |
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.)
|