wbmw,
What makes you so sure that Microsoft will release a client version of Windows for x86-64?
Dave Cutler: amd.com
I believe they've only committed to a version of .NET server. The next client Windows release isn't until late 2004, and they aren't likely to release anything out of cycle.
This is a problem I have posted about several times in the past. I don't have an answer to it. The lack of "vehicle" with which x86-64 can be bundled is a big concern, IMO. On the other hand, if MSFT decides to release it separately, that would be a huge coup for AMD, since there would be a lot of publicity that would go with it.
I think you have that backwards. IA-64 is supposed to be a 64-bit ISA without the "baggage" of x86.
Of course I meant baggage, or burden of switching to new instruction set, and all the costs that go with that.
Also, are you comparing Hammer to the .13u version of Itanium 2, on schedule to be launched around the same time?
I don't know where the companies will actually be at which time, but both will be excellent database performers.
Not based on your logic.
My quote that you took was part of my comment about the pundits talking about 4GB barrier. I think the % of desktop users needing > 4 GB will be tiny, but the media pundits will be talking about it. And it will actually be possible to purchase and install > 4 GB memory at a down to earth price in 2 years or so, which will make it more of an issue.
Unlike AMD's 32-bit processors, Intel's can address up to 64GB of memory, and there server products have been capable of >4GB for years already.
I believe Athlon K7 can address virtual memory > 4GB, but I don't believe any chipset released so far supports it (and probably never will, with Hammer on the horizon). As far as 32bit processors accessing > 4 GB, it is ugly.
Joe |
