Microsoft will have to offer 2 completely different versions of Windows and make sure that they are compatible between Windows platforms? Developers will have to code drivers and applications for two different Windows platforms? Users will somehow be lead to believe that there will be 100% compatibility between those two different Windows platforms? I doubt it... More upgrade marketing, more enduser confusion. Not to mention that Microsoft has not officially pledged support for AMDs 64-bit chips. I forward your attention to "Cowards" post. Seeing a 64-bit Windows desktop probably *won't* happen until around the end of the decade.
--
Ed M.</strong><hr></blockquote>
It's not the first time MS has had windows for more than one platform in development, production and use at the same time. WinNT 4 was out for x86, Alpha and a third (PPC, I think). They dropped the last two because of lack of sales.
The thing about x86-64 is that developers and users don't have to worry about 2 platforms. It totally capable of operating with x86-32 so developers can target both or just -32. What do you think happened with the 386? Windows 3.1 brought in win32 and used it on 386s(486s) but not 286s... No probs (kinda... it is windows after all).
MS will release a version of windows for x86-64 this year or next - guaranteed. Apple needs to count on this.
It's not the first time MS has had windows for more than one platform in development, production and use at the same time. WinNT 4 was out for x86, Alpha and a third (PPC, I think). They dropped the last two because of lack of sales.</strong><hr></blockquote>
And NT for Alpha didn't sell because it ran like a dog (on an Alpha!) and applications had to be recompiled to run on it - which few were.
The result was that anyone interested in ponying up for Alpha hardware ran OSF/1 (now Tru64 UNIX) or OpenVMS.
Itanic is running into the same wall that Alpha did with respect to Windows, except that IA-64 wants to be Alpha in its wildest fantasies.
[quote]<strong>The thing about x86-64 is that developers and users don't have to worry about 2 platforms.</strong><hr></blockquote>
The thing about MS releasing Server 2003 for IA-64 is that developers and users will have to worry about at least two totally different platforms: x86-64 and IA-64. AMD's 64-bit extensions add another sorta-platform (and IA-64's emulation of x86 is currently too weak to consider). This is not pretty. The fact that Windows has nothing like OS X's bundles makes it less pretty. Uglifying things further is the spectre of NT for Alpha. That OS is not half as portable as MS has bragged that it would be.
[quote]<strong>MS will release a version of windows for x86-64 this year or next - guaranteed. Apple needs to count on this.</strong><hr></blockquote>
Most likely, but if not then AMD might find itself riding the Linux wave, and the Linux wave might get a boost from general disgust with Itanium 2 - not because it performs poorly (running IA-64 apps), but because it really is a whole new platform that effectively orphans a huge legacy of IA-32 apps. Linux, on the other hand, should be able to use the Opteron to scale up to 64 bit almost (but not quite) as easily and almost (but not quite) as efficiently as OS X will jump from 32-bit PPC to 64-bit PPC.
As many years as I've waited for that sucktacular ISA to die a well-deserved death, part of me is gleefully happy at the prospect of Yet Another Hack to Prolong x86 killing Intel's attempt to get away from it. I think of it as karmic payback.
<strong>But once Apple released OS X with some 64-bit optimized code, it'll fly. My bets are that that will be 10.3, released when the first 970s are announced.</strong><hr></blockquote>
As I said, there isn't anything in particular about the OS or 64-bitness that will make it "fly". The important step will be simply recompiling with a version of GCC that is POWER4/970 aware so that it schedules according to the new processor's preferences rather than the G4's preferences. It remains to be seen how much of a difference that makes.
Comments
<strong>
Microsoft will have to offer 2 completely different versions of Windows and make sure that they are compatible between Windows platforms? Developers will have to code drivers and applications for two different Windows platforms? Users will somehow be lead to believe that there will be 100% compatibility between those two different Windows platforms? I doubt it... More upgrade marketing, more enduser confusion. Not to mention that Microsoft has not officially pledged support for AMDs 64-bit chips. I forward your attention to "Cowards" post. Seeing a 64-bit Windows desktop probably *won't* happen until around the end of the decade.
--
Ed M.</strong><hr></blockquote>
It's not the first time MS has had windows for more than one platform in development, production and use at the same time. WinNT 4 was out for x86, Alpha and a third (PPC, I think). They dropped the last two because of lack of sales.
The thing about x86-64 is that developers and users don't have to worry about 2 platforms. It totally capable of operating with x86-32 so developers can target both or just -32. What do you think happened with the 386? Windows 3.1 brought in win32 and used it on 386s(486s) but not 286s... No probs (kinda... it is windows after all).
MS will release a version of windows for x86-64 this year or next - guaranteed. Apple needs to count on this.
MM
<strong>
It's not the first time MS has had windows for more than one platform in development, production and use at the same time. WinNT 4 was out for x86, Alpha and a third (PPC, I think). They dropped the last two because of lack of sales.</strong><hr></blockquote>
And NT for Alpha didn't sell because it ran like a dog (on an Alpha!) and applications had to be recompiled to run on it - which few were.
The result was that anyone interested in ponying up for Alpha hardware ran OSF/1 (now Tru64 UNIX) or OpenVMS.
Itanic is running into the same wall that Alpha did with respect to Windows, except that IA-64 wants to be Alpha in its wildest fantasies.
[quote]<strong>The thing about x86-64 is that developers and users don't have to worry about 2 platforms.</strong><hr></blockquote>
The thing about MS releasing Server 2003 for IA-64 is that developers and users will have to worry about at least two totally different platforms: x86-64 and IA-64. AMD's 64-bit extensions add another sorta-platform (and IA-64's emulation of x86 is currently too weak to consider). This is not pretty. The fact that Windows has nothing like OS X's bundles makes it less pretty. Uglifying things further is the spectre of NT for Alpha. That OS is not half as portable as MS has bragged that it would be.
[quote]<strong>MS will release a version of windows for x86-64 this year or next - guaranteed. Apple needs to count on this.</strong><hr></blockquote>
Most likely, but if not then AMD might find itself riding the Linux wave, and the Linux wave might get a boost from general disgust with Itanium 2 - not because it performs poorly (running IA-64 apps), but because it really is a whole new platform that effectively orphans a huge legacy of IA-32 apps. Linux, on the other hand, should be able to use the Opteron to scale up to 64 bit almost (but not quite) as easily and almost (but not quite) as efficiently as OS X will jump from 32-bit PPC to 64-bit PPC.
As many years as I've waited for that sucktacular ISA to die a well-deserved death, part of me is gleefully happy at the prospect of Yet Another Hack to Prolong x86 killing Intel's attempt to get away from it. I think of it as karmic payback.
Not to mention that Microsoft has not officially pledged support for AMDs 64-bit chips. <hr></blockquote>
<a href="http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543~19906,00.html"; target="_blank">Relevant Press Release </a>
[quote]
I forward your attention to "Cowards" post. Seeing a 64-bit Windows desktop probably *won't* happen until around the end of the decade.
--
Ed M.<hr></blockquote>
Seeing how I've already seen and used that 64-bit Windows desktop I'd have to disagree with your assessment.
<strong>But once Apple released OS X with some 64-bit optimized code, it'll fly. My bets are that that will be 10.3, released when the first 970s are announced.</strong><hr></blockquote>
As I said, there isn't anything in particular about the OS or 64-bitness that will make it "fly". The important step will be simply recompiling with a version of GCC that is POWER4/970 aware so that it schedules according to the new processor's preferences rather than the G4's preferences. It remains to be seen how much of a difference that makes.