Xebec
Well-known member
Asianometry has posted a "post-mortem" on Itanium, covering the full history from beginning to end:
This video is pretty thorough, and a few things surprised me as different than my memory on the product*. Jon (from Asianometry) quotes high quality sources such as the oral histories of Bob Caldwell and other engineers, and sprinkles in quotes from various DEC / Compaq / Intel leaders, and the Microprocessor Report to add perspective. There's a good balance of technical ("magic compilers") and non-technical topics (i.e. the formation of Intel's second major CPU engineering team) here, too.
Anyway, if you were ever curious about the genesis of the product, and it's full effects on the industry (not always bad -- Itanium did effectively displace several architectures that competed with Intel, but laid the foundation for a very resurgent AMD), this is a great "historical" video.
..
*For example, he used sources that indicated Itanium was never intended to replace x86 -- but to conquer high end servers, where I always had the impression it was to eventually replace x86. The strategy I recall (incorrectly?) was that Intel would push Itanium native applications to new heights, where emulation of old apps would be "good enough" to allow a transition over a long period.
Last thought - Jon also mentioned some of the challenges and delays around both Merced and McKinley -- Merced mostly just ran late, and McKinley had to strip a lot of important features just to fit in a transistor budget. I would like to hear more about the engineering and technical challenges of the first 3-4 Itanium revisions in the future.
This video is pretty thorough, and a few things surprised me as different than my memory on the product*. Jon (from Asianometry) quotes high quality sources such as the oral histories of Bob Caldwell and other engineers, and sprinkles in quotes from various DEC / Compaq / Intel leaders, and the Microprocessor Report to add perspective. There's a good balance of technical ("magic compilers") and non-technical topics (i.e. the formation of Intel's second major CPU engineering team) here, too.
Anyway, if you were ever curious about the genesis of the product, and it's full effects on the industry (not always bad -- Itanium did effectively displace several architectures that competed with Intel, but laid the foundation for a very resurgent AMD), this is a great "historical" video.
..
*For example, he used sources that indicated Itanium was never intended to replace x86 -- but to conquer high end servers, where I always had the impression it was to eventually replace x86. The strategy I recall (incorrectly?) was that Intel would push Itanium native applications to new heights, where emulation of old apps would be "good enough" to allow a transition over a long period.
Last thought - Jon also mentioned some of the challenges and delays around both Merced and McKinley -- Merced mostly just ran late, and McKinley had to strip a lot of important features just to fit in a transistor budget. I would like to hear more about the engineering and technical challenges of the first 3-4 Itanium revisions in the future.