[Developers] Re: [Flesh-cvs] DEVELOPMENT CVS "Cactus/doc ReleaseNotes"
goodale at cct.lsu.edu
Tue Nov 2 18:26:16 CST 2004
Interesting. Would you be able to test/give us access to test, on one of
these AMD boxes for the next beta release ?
On Tue, 2 Nov 2004, John Shalf wrote:
> Well one thing that is interesting is that the Intel beta 8.1 compilers for
> the EMT64 generate code that outright refuses to run on our AMD boxes. Intel
> support told us that this was to protect us from the subtle differences
> between the two architectures, but wasn't able to provide any depth to the
> description. It must be inserting some code to check some privileged/hidden
> register on the nocona chips because I can't imagine how it would figure it
> out otherwise. The differing instructions are only invoked for OS activites
> (its not something that makes sense to use in user space).
> On Nov 2, 2004, at 4:11 PM, Tom Goodale wrote:
>> I thought there were some subtle differences between them ? So as the box
>> we did test on was an intel one, it may be best to be precise.
>> On Wed, 3 Nov 2004, Jonathan Thornburg wrote:
>>> Hi, Tom,
>>> In the beta15 release notes just committed to CVS, you wrote
>>>> - We have tested Cactus on three new architectures, the SGI Altix,
>>>> Intel's new Extended Memory 64 Technology, and Microsoft's XBox,
>>>> all running Linux.
>>> Should we maybe mention that this is often called "AMD64"?
>>> Notably, that's what http://www.x86-64.org was calling it long before
>>> intel came up with the name "EMT64".
>>> As another exmaple, http://www.redhat.com/software/rhel/es/
>>> calls it "AMD64/EMT64". I would suggest mentioning all three names:
>>> x86-64, AMD64, and EMT64.
>> Developers mailing list
>> Developers at cactuscode.org
> Developers mailing list
> Developers at cactuscode.org
More information about the Developers