rec.autos.simulators

I briefly skimmed the article, it's rather misleading in the fact
that the Alpha is essentially being discarded. That's not the case.
The Alpha has been somewhat stagnant since the Digital/Compaq merger,
the cost to R&D and produce the Alpha put a halt on further design
and implementation.

Teaming with Intel has provided the opportunity to lower R&D costs,
produce the Alpha based core much cheaper than Compaq ever could and
to aggressively market an Enterprise solution to take on the likes
of IBM and Sun. Alpha at the moment is floundering against them because
of the stagnant design of the Alpha.

Compaq is relying on an Intel hybrid Alpha to be able to push a hell
of a lot more Alphas out there and attempt to takeover the Enterprise
space. The Compaq Intel server range competes exceptionally well
against IBM and HP for instance.

Depends on the server. I already work on servers that have 3 busses, 8
processors and 16GB or RAM. Utilising separate busses have been in Intel
servers for a number of years now.

The lower end of the market has had multiple bus architecture for
quite awhile. Dual processor, 8GB servers.

It however doesn't solve the problem where most servers suffer. Disk
transfer and network bandwidth.

Don't get me wrong, I have nothing against AMD and my next machine is
likely to be one. We can both argue how narrowminded it is for a company
purchasing an Intel based server also purchase Intel based desktops and
not AMD based, but that's the reality. AMD could have a nice slice of
the consumer market but that's not where the money is. They have a lot
of work to do to break into the commercial market. The only AMD PC's and
servers you'll likely see in an office is ones where the IT manager is
a rabid AMD fan.

Cheers,
Rod.

Processor cost is almost negligable when it comes to costing a server.
The few hundred dollar saving isn't neccessarily going to be seen
when you price disk and memory.

My argument is this;

- Consumer market is only a small slice of the pie when it comes to
  processor sales.

- AMD's market penetration on small to medium level servers in the
  commercial marketplace is almost insignificant.

- Large vendors are marketing AMD to the home/consumer market but not
  as aggressively to the commercial market as they do 'Intel Inside'
  boxes.

- An Intel server will generally mean Intel clients on the desktop.

- Intel now have access to the algorithms that make the DEC Alpha chip
  what it is. I'm led to believe this is the single most hurdle that
  Intel just couldn't conquer... and when they came close, DEC was
  slapping a copywright infringement on them. ;)

- Intel has provided the technology for 4 proccessor capable boxes since
  the MMX and 8 processor capable boxes since the P3 Xeons. Years of
  proven design in what the market wants, throwing more CPU's in the
  one box.

- CPU bandwidth is rarely the bottleneck on a properly spec'ed machine.

What AMD have done is nothing short of remarkable, greater CPU speed is
going to be great for us gamers, but it's not neccessarily going to push
processor sales up where they really need to be in the commercial market
until they can start competing across the whole range of Intel based
products.

Either way, at the moment we're still in a win-win situation, and it's
likely at the moment that my next machine will be an AMD.

Cheers,
Rod.

I asked because I honestly didn't know. ;) I thought dual capable
AMD's were around for awhile.

8.

Not neccessarily. On a vanilla deskop, certainly. On a mid-range
server, not neccessarily.

Have they released a brief on the design of their memory subsystem?
Wouldn't mind taking a look at it.

Processor speed isn't neccessarily the key to a server. Show me a
server that doesn't spend the majority of its CPU time peaking at
a maximum 10% processor utilisation.

Disk and network are your bottlenecks in an exceptionally high
number of cases in a properly spec'ed multiprocessor environment.
The cost of improving disk and network speed makes the difference
between an Intel chip and an AMD chip look like spare change.
Intel already have thumbs firmly entrenched in both of those pies
as well.

If you can get the 'Big 4 - Compaq, IBM, HP, Dell' to start
shipping these in any sort of volume, then AMD have a good
platform to scare the heebeejeebee's out of Intel. Until then,
they'll never reach the market penetration they desire (in my
humble opinion, anyway).

Cheers,
Rod.

Multiple busses, etc. have been around for a while, but mainly in the
Enterprise applications, and they (Intel) enjoy a no-real-competition status
in there for now (hence the pricing).

Currently specing an 4-8 processor server right now for my DB-server, and
that leaves me with... well.... Intel.

Although all my desktops and utility servers will be AMD. I'm really looking
forward to enjoying some pricing wars in the dual processor marketplace so I
can keep costs down... and will prolly go with AMD on those 'just because'.

Either way... agree with you again... the competition/price-war can continue
for another 30 years as far as my desires go... we the consumer do win in
the end.

BTW... on teh topic of disk access times/net-bandwidth issues.... you played
around with any solid-state storage configs? They look pretty nice... but
paaaaaarrrrrriiiiiiiiiccccyyyyyy! Although... what else would you expect
from what might otherwise be simply looked at as a giant waffer box of RAM.

Cheers,

Schumi

| says...
| > That's where I heard it first (Alpha discontinuation). I was always a
fan of
| > the Alpha technology (which was 10 years ahead of its time). Also note
that
| > the article states that it is a discontinuation by Compaq, and that
Compaq
| > are going to embrace the Itanium chipset in the future (because of
existing
| > multi-processor capabilities). But I suspect that Compaq will be looking
at
| > AMD shortly, when they see the new architectures in action.
|
| I briefly skimmed the article, it's rather misleading in the fact
| that the Alpha is essentially being discarded. That's not the case.
| The Alpha has been somewhat stagnant since the Digital/Compaq merger,
| the cost to R&D and produce the Alpha put a halt on further design
| and implementation.
|
| Teaming with Intel has provided the opportunity to lower R&D costs,
| produce the Alpha based core much cheaper than Compaq ever could and
| to aggressively market an Enterprise solution to take on the likes
| of IBM and Sun. Alpha at the moment is floundering against them because
| of the stagnant design of the Alpha.
|
| Compaq is relying on an Intel hybrid Alpha to be able to push a hell
| of a lot more Alphas out there and attempt to takeover the Enterprise
| space. The Compaq Intel server range competes exceptionally well
| against IBM and HP for instance.
|
| > AMD is just now enterring teh Server Market, but the glowing difference
is
| > that the AMD chipsets utilize seperate busses int heir architecture. As
| > ooposed to a single bus in teh Intel implementations.
|
| > So you'll have 2 64-bit buses running in tandem, and thus removing many
of
| > the bottlenecks.
|
| Depends on the server. I already work on servers that have 3 busses, 8
| processors and 16GB or RAM. Utilising separate busses have been in Intel
| servers for a number of years now.
|
| The lower end of the market has had multiple bus architecture for
| quite awhile. Dual processor, 8GB servers.
|
| It however doesn't solve the problem where most servers suffer. Disk
| transfer and network bandwidth.
|
| Don't get me wrong, I have nothing against AMD and my next machine is
| likely to be one. We can both argue how narrowminded it is for a company
| purchasing an Intel based server also purchase Intel based desktops and
| not AMD based, but that's the reality. AMD could have a nice slice of
| the consumer market but that's not where the money is. They have a lot
| of work to do to break into the commercial market. The only AMD PC's and
| servers you'll likely see in an office is ones where the IT manager is
| a rabid AMD fan.
|
| Cheers,
| Rod.

--
Alastair Ingram
www.saxlessons.com
ICQ#108243828

I agree with you wholeheartedly Rod.

Multiple busses, etc. have been around for a while, but mainly in the
Enterprise applications, and they (Intel) enjoy a no-real-competition status
in there for now (hence the pricing).

Currently specing an 4-8 processor server right now for my DB-server, and
that leaves me with... well.... Intel.

Although all my desktops and utility servers will be AMD. I'm really looking
forward to enjoying some pricing wars in the dual processor marketplace so I
can keep costs down... and will prolly go with AMD on those 'just because'.

Either way... agree with you again... the competition/price-war can continue
for another 30 years as far as my desires go... we the consumer do win in
the end.

BTW... on teh topic of disk access times/net-bandwidth issues.... you played
around with any solid-state storage configs? They look pretty nice... but
paaaaaarrrrrriiiiiiiiiccccyyyyyy! Although... what else would you expect
from what might otherwise be simply looked at as a giant waffer box of RAM.

Cheers,

Schumi

| says...
| > That's where I heard it first (Alpha discontinuation). I was always a
fan of
| > the Alpha technology (which was 10 years ahead of its time). Also note
that
| > the article states that it is a discontinuation by Compaq, and that
Compaq
| > are going to embrace the Itanium chipset in the future (because of
existing
| > multi-processor capabilities). But I suspect that Compaq will be looking
at
| > AMD shortly, when they see the new architectures in action.
|
| I briefly skimmed the article, it's rather misleading in the fact
| that the Alpha is essentially being discarded. That's not the case.
| The Alpha has been somewhat stagnant since the Digital/Compaq merger,
| the cost to R&D and produce the Alpha put a halt on further design
| and implementation.
|
| Teaming with Intel has provided the opportunity to lower R&D costs,
| produce the Alpha based core much cheaper than Compaq ever could and
| to aggressively market an Enterprise solution to take on the likes
| of IBM and Sun. Alpha at the moment is floundering against them because
| of the stagnant design of the Alpha.
|
| Compaq is relying on an Intel hybrid Alpha to be able to push a hell
| of a lot more Alphas out there and attempt to takeover the Enterprise
| space. The Compaq Intel server range competes exceptionally well
| against IBM and HP for instance.
|
| > AMD is just now enterring teh Server Market, but the glowing difference
is
| > that the AMD chipsets utilize seperate busses int heir architecture. As
| > ooposed to a single bus in teh Intel implementations.
|
| > So you'll have 2 64-bit buses running in tandem, and thus removing many
of
| > the bottlenecks.
|
| Depends on the server. I already work on servers that have 3 busses, 8
| processors and 16GB or RAM. Utilising separate busses have been in Intel
| servers for a number of years now.
|
| The lower end of the market has had multiple bus architecture for
| quite awhile. Dual processor, 8GB servers.
|
| It however doesn't solve the problem where most servers suffer. Disk
| transfer and network bandwidth.
|
| Don't get me wrong, I have nothing against AMD and my next machine is
| likely to be one. We can both argue how narrowminded it is for a company
| purchasing an Intel based server also purchase Intel based desktops and
| not AMD based, but that's the reality. AMD could have a nice slice of
| the consumer market but that's not where the money is. They have a lot
| of work to do to break into the commercial market. The only AMD PC's and
| servers you'll likely see in an office is ones where the IT manager is
| a rabid AMD fan.
|
| Cheers,
| Rod.

One could assume that end of the market you're competing against
Alpha, IBM, Sun, i.e different processor families all-together. But
generally a pricier, speedier and scalable alternative.

Not as yet. The market where I am have only just begun to embrace fibre
on SCSI disks in large storageworks boxes. I can't see too many Intel
configs that would run solid state at its current prices. You could
put together a solution with greater redundancy and higher availability
with a few bucks to spare at the cost of some performance. But solid
state seems to be marketed at the enterprise market that Intel can not
yet reach.

Cheers,
Rod.

AMD is a godsend to the home user. It's difficult to say where Intel's
technology would be without AMD pushing the envelope. I'm not yet
convinced with AMD's potential in the office space, but it
certainly can't do any harm. ;)

Cheers,
Rod.