Archived from groups: comp.sys.ibm.pc.games.action (
More info?)
On Mon, 28 Feb 2005 15:04:18 +0100, Mean_Chlorine
<mike_noren2002@NOSPAMyahoo.co.uk> wrote:
>Thusly john.dsl@verizon.net (John Lewis) Spake Unto All:
>
>>939 is AMDs forseeable-future pin-out for desktop applications
>>both single and dual-core.
>
>I allow myself to be skeptical about that. AMD is very much into
>planned obsolescense through socket change, and has a habit of not
>informing anyone of changes until they're imminent, so I'm guessing
>"the forseeable future" is one to two years.
>
>>The AMD desktop dual-core processors will only be available in
>>939-pin form and <will retrofit> into the current 939-socket boards
>>with just a BIOS change.
>
>Oh I'd like to believe that, it would thrill and surprise me if
>there's not a big "BUT..." somewhere.
>Like that those processors require more juice, or less juice, or don't
>support present-gen memory controllers.
AMD 64-bit has a built-in memory controller. The single reason for the
easy single/dual compatibility - to everything external, the processor
looks the same. A truly smart and forward-looking architectural
decision. Not so with Intel's external memory controllers - hence they
will need to update/kludge chip-sets if they want single/dual
compatibility on the same MB - and I have no idea whether that is
their intent with the 945 chip-set...........
>Something, *anything*, that
>mean they can not be used with present motherboards (or at least are
>severely crippled if they can).
>
Engineering samples of the AMD dual-core has already been tried out in
the Asus A8N-SLI. Works just fine, according to the reports. Built on
the 90nm process and run at a slightly-lower clock-rate than the
single-core; no problem with power/heat.
However, if you are an overclocker-- do not rush out and buy the
current A8N-SLI... see today's review by Anand on the current crop of
SLI motherboards.
>I use and recommend AMD. I just don't like their tendency for secrecy
>and the way they hinder upgrades by surprise-changing sockets at short
>intervals.
>
Have you observed Intel recently? Why shift from 478 to 775-pin at all
for the 32-bit single core, since they were planning for
near-simultaneous release of 64-bit and dual core. Why not delay
pin-out/MB gyrations until the full 64-bit release and architecture
full single-/dual-core/32-bit/64-bit compatibility into that
CPU-design, pin-out and associated chip-sets ?? Both the current
775-pin 32-bit Prescott and their 915/925 motherboards are soon-to-be
orphans, after far less than a year of production. Compare with the
life-time of the 478-pin P4 family and the associated chip-sets.
BTW, up to this point I have only had Intel processors in all my
machines and systems that I have built for others. However,
Intel's Prescott system-architectural and marketing stupidity plus
their arrogance with regard to the 64-bit roll-out, taken together
with AMDs faultless 64-bit execution in both desk-top and X86-servers
has made me ( like many others) actively consider the alternatives.
And on the server-side, I note Anand's glowing review of the
4x Opteron 3U-high baby-server from Sun --- into which
4 dual-core Opterons can be instantly swapped when they
are production-available.............hah !! At ~ $20,000
a pop with instant dual-core upgrade potential, Sun should
sell a whole lot of those..........
John Lewis.
Facebook
Twitter
Instagram