[Info-vax] Intel: Kittson to be socket-compatible with 9300 and 9500

[John Wallace]

On Feb 4, 7:30 pm, Keith Parris <keithparris_deletet... at yahoo.com>
wrote:
> During development of the Tukwila (9300) chip, Intel and HP adjusted
> plans in response to market directions and accepted the extra
> development time required to redesign Tukwila (and the planned HP
> systems) to use DDR3 memory instead of the previously-planned DDR2. This
> provided better performance for the chip and the systems designed around it.
>
> In 2010, Intel introduced its common platform strategy that allowed
> Intel Itanium and Xeon processors to utilize common platform ingredients
> including chipsets, interconnects and memory. Xeon would gain RAS
> features from Itanium and Itanium would gain from Xeon efficiencies of
> scale due to high volumes.
>
> At the time the HP/Oracle lawsuit exposed details of HP's strategy from
> June 2010 code-named "Kinetic"
> (http://h20341.www2.hp.com/integrity/downloads/about-kinetic.pdf), it
> was seen that plans at the time were for the Kittson chip, the Itanium
> chip expected 2-3 years after Poulson, to have a socket compatible with
> that of Xeon. But by the time of that disclosure, during the summer of
> 2012, Intel had become coy and was no longer willing to comment about
> plans for Kittson socket designs.
>
> At the November 8, 2012 announcement of the Poulson (9500) processor
> (http://newsroom.intel.com/community/intel_newsroom/blog/2012/11/08/ne...),
> Intel again touted its Modular Development Model which shares
> silicon-level design elements between Itanium and Xeon for logic other
> than the processor cores, and said plans were for Kittson to have a
> socket compatible with Xeon. But then last Thursday, Intel announced
> that Kittson would instead use a socket compatible with the Tukwila and
> Poulson processors, and be released in the same 32 nm process as
> Poulson. Using the same socket and process will undoubtedly allow better
> time-to-market for the Kittson processor and related system designs from
> HP, plus better potential investment protection for customers owning
> Integrity systems with Tukwila or Poulson processors. Intel did keep
> open the option of using a Xeon socket for a subsequent Itanium
> processor. The Intel announcement is athttp://www.intel.com/content/www/us/en/processors/itanium/itanium-kit....
> Intel also provided an updated customer letter which covers Kittson athttps://pit.houston.hp.com/PIT/ESSN/Document_Storage/collateral/Intel...

Tim Pricket Morgan at The Register now says Intel are backtracking on
the "common socket" story.

http://www.theregister.co.uk/2013/02/08/intel_kills_itanium_xeon_convergence_and_kittson/

Not sure who's the least credible in this picture.