> is because there is a penalty for executing 32-bit code, one which is
There is none (if you use the 32bit subset of the AMD64 architecture).
But there is a penalty for using the x86 architecture instead of the
amd64 architecture.
This penalty is not specific to the Athlon64/Opteron/younameit, but to all
known x86 implementation: it's just the the x86 architecture has few
registers and this limitation was partly lifted in the amd64 architecture.
So the "penalty" you're talking about, is there because the amd64
architecture did more than extend pointers and int to 64bit, it also fixed
a few problems in the x86 architecture. It has nothing to do with whether
or not the amd64 is a "true 64 bit architecture", whatever that may mean.
Stefan
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
