Path: utzoo!attcan!uunet!lll-winken!maddog!brooks
From: brooks@maddog.llnl.gov (Eugene Brooks)
Newsgroups: comp.arch
Subject: Re: Shared Memory (was: ETA-10: CMOS or ECL?)
Keywords: Cray Shared Memory
Message-ID: <13580@lll-winken.llnl.gov>
Date: 1 Nov 88 17:34:34 GMT
References: <13529@lll-winken.llnl.gov> <850@super.ORG>
Sender: usenet@lll-winken.llnl.gov
Reply-To: brooks@maddog.UUCP (Eugene Brooks)
Organization: Lawrence Livermore National Laboratory
Lines: 9

In article <850@super.ORG> laurel@super.UUCP (Michael Tighe) writes:
>I would not call Cray's version of shared memory, real honest to God 
>shared memory. There are many things it can't do, that other systems can,
>such as shared text and the like.
The Cray hardware supports seperate text and data segments, whether
the text is "shared" or not is a feature of the operating system and
not the hardware.  It would be nice if there were more than one data
segment, so that private memory with staticly allocated addresses
could be supported, provided that it did not make the machine go
slower.