3 ms·

A History of Modern 64-bit Computing (2007) [pdf]

29 points by Aloha 11y ago

chx 11y ago
If you found this PDF interesting then also read this 2009 article from John C. Dvorak http://www.pcmag.com/article2/0,2817,2339629,00.asp http://www.pcmag.com/article2/0,2817,2339629,00.asp
- eru 11y ago
  The PDF was interesting, but the Dvorak article linked was meh.
fulafel 11y ago
This mentions only in passing the 64-bit RISCs that preceded AMD64 & Itanium. Also some outright fiction like "IA64 provided unmatched performance on scale-up workloads". I guess the Microsoft authors don't want to talk about shipping a 32-bit NT on those while the competition had all 64-bit sw stacks :) Very interesting read still.
- Aloha 11y ago
  I'd argue that there was no true demand for 64 bit computing prior to 2001-2002, if you look at say the Sun Ultra series of workstations, most couldn't go beyond 4 gb of ram per processor, many much less, the same appears to be the case for DEC. The exception largely was SGI. Even then, the markets Microsoft was selling NT into, wouldn't have benefitted much from the extra memory (or the greater floating point performance) allowed by 64-Bit processors. Microsoft has never been in the scientific computing part of the business that SUN, DEC, and SGI were playing in.
  - fulafel 11y ago
    Sure, workstations didn't need 64-bit at that time. But servers did, and at that time Windows was still trying to prove its credibility as a server OS vs the incumbent Sun/HP/IBM/etc platforms. Another field where 64-bit was an even more urgent requirement was HPC. At the time HPC was more focused on big iron than PC clusters. Here's a classic post from John Mashey outlining the rationale for 64-bit in the early 90s: http://yarchive.net/comp/64bit.html http://yarchive.net/comp/64bit.html Also remember how long and painful the 16-to-32 transition was on Microsoft's platforms. They had every reason to get their platform and running up many years before customers needed the 64-bit apps.
    - Aloha 11y ago
      Yeah, that's more or less the argument I was making - (though I seem to recall that it wasn't until the late 90's that the sun user land was 64bit, it was all still 32bit) I don't consider Windows Server a really good general purpose server system, its excellent at some workloads though. IA-64 was a mistake, it has hurt everyone who touched it, I believe it killed VMS (beyond the crazy way VMS is licensed) (and did no favors for HP-UX either.) - it is and was too hard to build a really efficient compiler for - thankfully microsofts porting effort to IA-64 got the codebase set up to go to AMD64. In many ways though, as much as I'd like to see other microarchitectures, x86 is the UNIX of hardware, it's good enough for most applications, and in the long run, good always beats perfect.

29 points by Aloha 11y ago

chx 11y ago

If you found this PDF interesting then also read this 2009 article from John C. Dvorak http://www.pcmag.com/article2/0,2817,2339629,00.asp http://www.pcmag.com/article2/0,2817,2339629,00.asp

eru 11y ago
The PDF was interesting, but the Dvorak article linked was meh.

fulafel 11y ago

This mentions only in passing the 64-bit RISCs that preceded AMD64 & Itanium. Also some outright fiction like "IA64 provided unmatched performance on scale-up workloads". I guess the Microsoft authors don't want to talk about shipping a 32-bit NT on those while the competition had all 64-bit sw stacks :) Very interesting read still.

Aloha 11y ago
I'd argue that there was no true demand for 64 bit computing prior to 2001-2002, if you look at say the Sun Ultra series of workstations, most couldn't go beyond 4 gb of ram per processor, many much less, the same appears to be the case for DEC. The exception largely was SGI. Even then, the markets Microsoft was selling NT into, wouldn't have benefitted much from the extra memory (or the greater floating point performance) allowed by 64-Bit processors. Microsoft has never been in the scientific computing part of the business that SUN, DEC, and SGI were playing in.
- fulafel 11y ago
  Sure, workstations didn't need 64-bit at that time. But servers did, and at that time Windows was still trying to prove its credibility as a server OS vs the incumbent Sun/HP/IBM/etc platforms. Another field where 64-bit was an even more urgent requirement was HPC. At the time HPC was more focused on big iron than PC clusters. Here's a classic post from John Mashey outlining the rationale for 64-bit in the early 90s: http://yarchive.net/comp/64bit.html http://yarchive.net/comp/64bit.html Also remember how long and painful the 16-to-32 transition was on Microsoft's platforms. They had every reason to get their platform and running up many years before customers needed the 64-bit apps.
  - Aloha 11y ago
    Yeah, that's more or less the argument I was making - (though I seem to recall that it wasn't until the late 90's that the sun user land was 64bit, it was all still 32bit) I don't consider Windows Server a really good general purpose server system, its excellent at some workloads though. IA-64 was a mistake, it has hurt everyone who touched it, I believe it killed VMS (beyond the crazy way VMS is licensed) (and did no favors for HP-UX either.) - it is and was too hard to build a really efficient compiler for - thankfully microsofts porting effort to IA-64 got the codebase set up to go to AMD64. In many ways though, as much as I'd like to see other microarchitectures, x86 is the UNIX of hardware, it's good enough for most applications, and in the long run, good always beats perfect.