3 ms·

Your causality doesn't seem to line up. As you tell it, the workstation CPUs were already peaking in terms of complexity and cost, when Intel and HP announce a

Your causality doesn't seem to line up. As you tell it, the workstation CPUs were already peaking in terms of complexity and cost, when Intel and HP announce a huge investment in the next generation of this class of CPU. This doesn't signal that the segment is economically unviable, and if you're an existing player with an existing strong design, it signals an opportunity to invest in your own next generation and meet Intel/HP head-on with a superior brand (for this segment) and a more plausible next-gen design. Intel and HP show up promising a clean sheet design in a decade doesn't seem like a time to roll up your market leading offering and go home.

jeffbee 5y ago
Nothing about this history lines up with facts. Intel killed the RISC workstation CPUs with the Pentium Pro, not with the Itanium. PRO/Engineer and SolidWorks had been ported to Windows NT at the very beginning of that operating system's availability, and the performance on x86 was already better than RISC workstations, at a tiny fraction of the cost, by late 1995.
- throwaway9870 5y ago
  The Pentium Pro and eventually the AMD64 chips equalled or bested the workstations on many tasks by the end of the 90's and early 2000s. But the Itanium processor distracted them from even competing. SGI for example had at least one high-end MIPS processor under development (iirc "the beast") and they stopped the project and later were trying to ship Itanium servers. HP threw the towel in and joined forces with Intel. Here is an article from 1999 on Sun booting SunOS on Itanium: https://www.zdnet.com/article/sun-boots-solaris-on-itanium-hardware-5000103621/ https://www.zdnet.com/article/sun-boots-solaris-on-itanium-h... Compaq stopped making Alpha servers and switched to Itanium: https://en.wikipedia.org/wiki/DEC_Alpha https://en.wikipedia.org/wiki/DEC_Alpha Itanium was a huge misdirection and distraction for everyone in the workstation business and basically put them in such a bad spot they failed to even compete with Intel x86 past 2000 or so in any real way.
cpleppert 5y ago
The Itanium wasn't just designed as a workstation CPU it was designed to be superior to existing high-end systems and then over time with volume be the eventual x86 replacement. The business case for itanium didn't work as a workstation cpu. >>Intel and HP show up promising a clean sheet design in a decade doesn't seem like a time to roll up your market leading offering and go home. They didn't roll up so much as have their existing market disappear. SGI, DEC and Compaq were dead. Sparc and Power gradually transitioned away from any workstation offering to servers.

jeffbee 5y ago

Nothing about this history lines up with facts. Intel killed the RISC workstation CPUs with the Pentium Pro, not with the Itanium. PRO/Engineer and SolidWorks had been ported to Windows NT at the very beginning of that operating system's availability, and the performance on x86 was already better than RISC workstations, at a tiny fraction of the cost, by late 1995.

throwaway9870 5y ago
The Pentium Pro and eventually the AMD64 chips equalled or bested the workstations on many tasks by the end of the 90's and early 2000s. But the Itanium processor distracted them from even competing. SGI for example had at least one high-end MIPS processor under development (iirc "the beast") and they stopped the project and later were trying to ship Itanium servers. HP threw the towel in and joined forces with Intel. Here is an article from 1999 on Sun booting SunOS on Itanium: https://www.zdnet.com/article/sun-boots-solaris-on-itanium-hardware-5000103621/ https://www.zdnet.com/article/sun-boots-solaris-on-itanium-h... Compaq stopped making Alpha servers and switched to Itanium: https://en.wikipedia.org/wiki/DEC_Alpha https://en.wikipedia.org/wiki/DEC_Alpha Itanium was a huge misdirection and distraction for everyone in the workstation business and basically put them in such a bad spot they failed to even compete with Intel x86 past 2000 or so in any real way.

cpleppert 5y ago

The Itanium wasn't just designed as a workstation CPU it was designed to be superior to existing high-end systems and then over time with volume be the eventual x86 replacement. The business case for itanium didn't work as a workstation cpu. >>Intel and HP show up promising a clean sheet design in a decade doesn't seem like a time to roll up your market leading offering and go home. They didn't roll up so much as have their existing market disappear. SGI, DEC and Compaq were dead. Sparc and Power gradually transitioned away from any workstation offering to servers.