Subj : x86 memory model (was: [.NET] Volatile Fields [correction]) To : comp.programming.threads From : David Hopwood Date : Tue Sep 20 2005 06:23 pm Leaving aside for the moment that you had made an incorrect statement specifically about processor consistency... Joe Seigh wrote: > David Hopwood wrote: >> >> : > > That's not the official x86 memory model. Andy Glew didn't make > clear whether he was talking about the x86 memory model as (unofficially) > specified or as implemented. And he was not speaking *for* Intel. We all seem to agree that the documentation is inadequate. So it's quite reasonable to ask for clarification from someone who should know. > You'd look very foolish trying to file a bug report to Intel for > behavior that was allowed by the documented memory model claiming > Andy Glew told you this was how it was supposed to work in comp.arch. Since it's not clear what the documentation does allow, I wouldn't feel foolish at all. -- David Hopwood .