Bill Henkel (noemail@yahoo.com) on 1/5/12 wrote:
---------------------------
[snip]
>Paul A. Clayton wrote:
>> Unfortunately, I do not think AMD is positioned to lead in the
>> re-introduction of off-chip cache. The learning curve for MCM fabrication
>> might excessively penalize their smaller production volume.
>
>The substrate in the package is just a small FR4 printed
>circuit board, not a fancy ceramic MCM. Almost every BGA
>package made today contains one. These substrates are
>fabricated by board shops, not semiconductor companies,
>and they are readily available today.

Do you happen to know what the cost, bandwidth, and
(perhaps even) power differences are between different
types of integration? While BGA technology is mature, I
was under the impression that silicon interposer was
progressing the learning curve rather quickly now and
provides much better bandwidth per area of interface at
projected reasonable incremental cost (once reasonably
mature).

By the way, would it be practical to use a small portion
of the silicon interposer 'chip' for I/O interfaces (using
a simple, full depreciated process technology--I receive
the impression that at least some I/O interfaces do not
benefit from smaller transistors)? Presumably the
"stepper" would be more of a 'leaper' (skipping over area
that is just for the interposer), though there might be
some advantage to simple amplification and switching for
the interposer function. The economics of such might
work out (though I am sceptical).