SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?


To: Tony Viola who wrote (44559)1/1/1999 10:54:00 PM
From: Maxwell  Read Replies (3) | Respond to of 1572158
 
Tony Viola:

<<Can you substitute CPU chips in for Xeons? Doesn't matter, even if AMD were to pull off the big upset, it doesn't mean the boys that make servers would go with AMD. Remember who has the infrastructure support, best FIT rates, RAS, yields and availability of large quantities of chips practically guaranteed. In the case of servers, these are much more important than a little more speed. Ask IBM, Compaq, Dell, HP, anyone. This isn't about overclocking chips or some 16 year old's benchmarks. The server boys have their own benchmarks.>>

To be able to sell servers to IBM or CPQ it must pass IBM's and CPQ's internal proprietary benchmarks not some garbage Winstone benchmark. IBM and CPQ will know which system is better. K7 will outrank Xeons. Who is talking on overclocking? The K7 is all raw speed. If AMD can pull off the biggest upset then my AMD investment will have met its objective.

Maxwell



To: Tony Viola who wrote (44559)1/2/1999 1:07:00 PM
From: Ali Chen  Read Replies (1) | Respond to of 1572158
 
Tony, your repetitive arguments are mostly nonsense.
<Remember who has the infrastructure support,
best FIT rates, RAS, yields and availability
of large quantities of chips practically guaranteed.>
A server manufacturers make a product - servers.
How the heck your "yield and availability of
large quantities" may affect the choice of
CPU? Do you know how good or bad are yields
for 21164, or MIPS, or PA-8200? How large
their quantities are? Does it prevent
server makers from successfully using these CPUs?
As soon as a CPU
manufacturer guarantees parameters and
availability, who cares? In theory it is
better to have a second source, but you
know what happened to this institution
when Intel announced monopoly...
And why would "large quantities" be a
factor in the server market, where a
SINGLE machine supposed to serve many
clients? Unless you want to built a
4-way Xeon into every light bulb to
control the dimmer...

If you mean under "availability" the
continuous supply of certain CPUs
for which the service infrasrtucture
has been established and is profitable
(for a server maker of course),
juggling with processor's lines (as
Intel does) is also not a good argument
for you. Compare: Socket8 PentiumPro ->
Slot1/PII -> Slo2/Xeon...

<infrastructure support> - I am not really
sure you understand what you are talking
about. If you mean the passive hardware,
then the K7 was specifically designed to
reuse all it - connectors, brackets, thermal
plates, retainers, heat sinks, etc.

If you are talking
about chipsets, I don't think the recent
stories with 4way-NX-Xeon combo are very
supportive for your argument. In any case
for a state-of-the-art server people
used to have proprietary solutions for
SMP on x86 processors.

After all, Maxwell was talking primarily
about uniprocessor machines on K7 vs Xeon.
You must be so afraid of K7 to penetrate
your lucrative server market that you are
responding to you own fears! And in a way
pleasurable for you... As far as
I know, AMD did not explicitly announced
their intent to go after server
market...yet:)