By: Fab Guy (fakeemailahker.delete@this.hfkah.com),
Room: Moderated Discussions
Dean Kent (dkent@realworldtech.com) on 9/27/04 wrote:
---------------------------
>As for the 'historical' stuff, I have given up hope expecting people to be rational
>when discussing Intel's intent WRT Itanium. I have been looking around for over
>2 years to meet Paul's challenge - asking on comp.arch, searching Google, etc. -
>and have come up with nothing that can be directly attributed to Intel spokespersons.
>There are a lot of *claims*, but all without names or direct quotes/documents.
>Seems like people would rather rely on hearsay, assumption and 'common wisdom'
>than any actual evidence because it is what they *want* to believe. Suffice it
>to say that it is really unknown whether Intel really thought Itanium could replace
>x86 on the desktop - at least within 10 to 20 years. I've linked to statements
>from Andy Grove and Gordon Moore in the past where they stated (in '97/'98) that
>x86 would be around for a long time to come...
>
>Regards,
>Dean
Dean,
A couple places to search are the obvious, http://www.siliconstrategies.com and http://www.eetimes.com. Just put in "Merced" for the search. The oldest articles only go back to 1998 - 1999 however.
Any big library, especially a university, should have EE Times newspapers on micro-film from the mid 1990's. I used to collect them from the early and mid 90's but I don't think I have any around anymore.
Both of those web sites indicate that Merced, McKinley, & Foster were intended to be used in both workstations and servers.
Some quotes from the article are below. My recollection from the mid 1990's is that Intel promoted Merced and its successors as being for servers and workstations. They made big claims on performance and that it would dominate all other mainstream RISC cpus in it's class for volume. However, Intel was very clear that it wouldn't replace the desktop x86 or even begin to think about it for 10 years after introdcution.
For what it's worth, I've been an engineer in the semiconductor industry for 8+ years, and spent a few years before that in a semiconductor research environment... so I've followed the business fairly closely and have been right in the middle of it at times.
------------------------------------------
Craig Barrett himself....
Intel demos first Merced silicon at developer forum
Alexander Wolfe
Silicon Strategies
08/31/1999, 3:49 PM ET
The device is implemented in 0.18-micron CMOS, but Barrett wouldn't say when the processor will come to market. "We want it to come out as a very aggressive forward-looking product," he said. "It's a server/workstation product."
------------------------------------------
Commentary by the author....
Intel's Itanium to roll out at 800 MHz
Will Wade
EE Times
02/09/2000, 9:04 AM ET
The Itanium is expected to be used in servers and workstations, and Intel will support the processor with a chip set that allows up to four of the chips to work together. OEMs can then pack these units together into workhorse systems with as many as 512 Itanium chips working together.
-----------------------------------------------
I'm not sure this one makes sense.....
Intel executives insist Merced will be volume product
Mark Hachman
Silicon Strategies
09/01/1999, 1:24 PM ET
"Why invest billions of dollars in a test vehicle?" he asked. "Let me go on record. Ron Curry of Intel said Merced is a not a development vehicle. Merced is a product vehicle."
In fact, Ron Curry, director of marketing for Intel's Microprocessor Products Group, claimed that the total number of Merced chips sampling next year will outstrip the production of all devices manufactured by rivals for the RISC-based processor market. The production version of Merced will also surpass, or at least be "pretty competitive" with the market's best-selling RISC chips, Curry added.
-------------------------------------
---------------------------
>As for the 'historical' stuff, I have given up hope expecting people to be rational
>when discussing Intel's intent WRT Itanium. I have been looking around for over
>2 years to meet Paul's challenge - asking on comp.arch, searching Google, etc. -
>and have come up with nothing that can be directly attributed to Intel spokespersons.
>There are a lot of *claims*, but all without names or direct quotes/documents.
>Seems like people would rather rely on hearsay, assumption and 'common wisdom'
>than any actual evidence because it is what they *want* to believe. Suffice it
>to say that it is really unknown whether Intel really thought Itanium could replace
>x86 on the desktop - at least within 10 to 20 years. I've linked to statements
>from Andy Grove and Gordon Moore in the past where they stated (in '97/'98) that
>x86 would be around for a long time to come...
>
>Regards,
>Dean
Dean,
A couple places to search are the obvious, http://www.siliconstrategies.com and http://www.eetimes.com. Just put in "Merced" for the search. The oldest articles only go back to 1998 - 1999 however.
Any big library, especially a university, should have EE Times newspapers on micro-film from the mid 1990's. I used to collect them from the early and mid 90's but I don't think I have any around anymore.
Both of those web sites indicate that Merced, McKinley, & Foster were intended to be used in both workstations and servers.
Some quotes from the article are below. My recollection from the mid 1990's is that Intel promoted Merced and its successors as being for servers and workstations. They made big claims on performance and that it would dominate all other mainstream RISC cpus in it's class for volume. However, Intel was very clear that it wouldn't replace the desktop x86 or even begin to think about it for 10 years after introdcution.
For what it's worth, I've been an engineer in the semiconductor industry for 8+ years, and spent a few years before that in a semiconductor research environment... so I've followed the business fairly closely and have been right in the middle of it at times.
------------------------------------------
Craig Barrett himself....
Intel demos first Merced silicon at developer forum
Alexander Wolfe
Silicon Strategies
08/31/1999, 3:49 PM ET
The device is implemented in 0.18-micron CMOS, but Barrett wouldn't say when the processor will come to market. "We want it to come out as a very aggressive forward-looking product," he said. "It's a server/workstation product."
------------------------------------------
Commentary by the author....
Intel's Itanium to roll out at 800 MHz
Will Wade
EE Times
02/09/2000, 9:04 AM ET
The Itanium is expected to be used in servers and workstations, and Intel will support the processor with a chip set that allows up to four of the chips to work together. OEMs can then pack these units together into workhorse systems with as many as 512 Itanium chips working together.
-----------------------------------------------
I'm not sure this one makes sense.....
Intel executives insist Merced will be volume product
Mark Hachman
Silicon Strategies
09/01/1999, 1:24 PM ET
"Why invest billions of dollars in a test vehicle?" he asked. "Let me go on record. Ron Curry of Intel said Merced is a not a development vehicle. Merced is a product vehicle."
In fact, Ron Curry, director of marketing for Intel's Microprocessor Products Group, claimed that the total number of Merced chips sampling next year will outstrip the production of all devices manufactured by rivals for the RISC-based processor market. The production version of Merced will also surpass, or at least be "pretty competitive" with the market's best-selling RISC chips, Curry added.
-------------------------------------