1.8 gHz Xeon vs. 2.8 gHz P4

I am trying to decide between a 1.8 gHz Xeon and a 2.8 gHz P4 for a
single CPU file/print/database server for a small office. So far I have
not found any clear explanation of the differences in or advantages of
on over the other.

Can anyone provide any information or references?

--
_Bill_

That should be "one over the other".

--
_Bill_

Bill wrote:

I am trying to decide between a 1.8 gHz Xeon and a 2.8 gHz P4 for a
single CPU file/print/database server for a small office. So far I
have not found any clear explanation of the differences in or
advantages of on over the other.

Can anyone provide any information or references?

Bill wrote:

That should be "one over the other".

I'd go with the cheaper one and instead spend your
money on some extra ram and faster hard drives.
That will provide you more performance for this
kind of thing (file/print/database server)
Eric

--
I am Locutus of Borg.
Your life as it has been is over.
From this moment on you will service... us!

Thanks. That is the consensus of some others I have asked as well.

--
_Bill_

Eric wrote:

Bill wrote:

That should be "one over the other".

I'd go with the cheaper one and instead spend your
money on some extra ram and faster hard drives.
That will provide you more performance for this
kind of thing (file/print/database server)
Eric

Bill wrote:
I am trying to decide between a 1.8 gHz Xeon and a 2.8 gHz P4 for a
single CPU file/print/database server for a small office. So far I have
not found any clear explanation of the differences in or advantages of
on over the other.

Can anyone provide any information or references?

The Xeon fits in a server mobo which supports ECC memory. It is
therefore more reliable, both in terms of MTTF and undetected data
errors. That may justify the small extra cost depending on the cost of
errors, particularly undetected ones.

The P4 is probably less expensive and slightly faster.

--
bill davidsen )
SBC/Prodigy Yorktown Heights NY data center
Project Leader, USENET news
http://newsgroups.news.prodigy.com

Bill Davidsen wrote:
Bill wrote:

I am trying to decide between a 1.8 gHz Xeon and a 2.8 gHz P4 for a
single CPU file/print/database server for a small office. So far I have
not found any clear explanation of the differences in or advantages of
on over the other.

Can anyone provide any information or references?

The Xeon fits in a server mobo which supports ECC memory. It is
therefore more reliable, both in terms of MTTF and undetected data
errors. That may justify the small extra cost depending on the cost of
errors, particularly undetected ones.

The P4 is probably less expensive and slightly faster.


The P4 will be more than slightly faster, I expect. Faster bus, larger
cache, some of the performance bottlenecks inside were fixed (it is
Northwood or Prescott compared to the Willamette-based Xeon).

The Xeon probably supports two processors, so you can put a second
1.8GHz Xeon in it. It is possible, but unlikely, that the Xeon allows
up to four processors.

The memory types probably differ a great deal. The Xeon probably uses
expensive Rambus RIMMs while the P4 uses inexpensive, commodity DDR
SDRAM DIMMs. It is more likely that the Xeon memory will be protected
from errors, as Bill said.

Alex
--
My words are my own. They represent no other; they belong to no other.
Don't read anything into them or you may be required to compensate me
for violation of copyright. (I do not speak for my employer.)

On Wed, 05 Jan 2005 08:16:00 -0500, Alex Johnson wrote:

Bill Davidsen wrote:
Bill wrote:

I am trying to decide between a 1.8 gHz Xeon and a 2.8 gHz P4 for a
single CPU file/print/database server for a small office. So far I have
not found any clear explanation of the differences in or advantages of
on over the other.

Can anyone provide any information or references?

The Xeon fits in a server mobo which supports ECC memory. It is
therefore more reliable, both in terms of MTTF and undetected data
errors. That may justify the small extra cost depending on the cost of
errors, particularly undetected ones.

The P4 is probably less expensive and slightly faster.


The P4 will be more than slightly faster, I expect. Faster bus, larger
cache, some of the performance bottlenecks inside were fixed (it is
Northwood or Prescott compared to the Willamette-based Xeon).

The Xeon probably supports two processors, so you can put a second
1.8GHz Xeon in it. It is possible, but unlikely, that the Xeon allows
up to four processors.

The memory types probably differ a great deal. The Xeon probably uses
expensive Rambus RIMMs while the P4 uses inexpensive, commodity DDR
SDRAM DIMMs. It is more likely that the Xeon memory will be protected
from errors, as Bill said.

I might have missed one, but I don't believe there were any commercially
available P4 Xeon chipsets that used RDRAM...

/daytripper

daytripper wrote:
On Wed, 05 Jan 2005 08:16:00 -0500, Alex Johnson wrote:


Bill Davidsen wrote:

Bill wrote:


I am trying to decide between a 1.8 gHz Xeon and a 2.8 gHz P4 for a
single CPU file/print/database server for a small office. So far I have
not found any clear explanation of the differences in or advantages of
on over the other.

Can anyone provide any information or references?


The Xeon fits in a server mobo which supports ECC memory. It is
therefore more reliable, both in terms of MTTF and undetected data
errors. That may justify the small extra cost depending on the cost of
errors, particularly undetected ones.

The P4 is probably less expensive and slightly faster.


The P4 will be more than slightly faster, I expect. Faster bus, larger
cache, some of the performance bottlenecks inside were fixed (it is
Northwood or Prescott compared to the Willamette-based Xeon).

The Xeon probably supports two processors, so you can put a second
1.8GHz Xeon in it. It is possible, but unlikely, that the Xeon allows
up to four processors.

The memory types probably differ a great deal. The Xeon probably uses
expensive Rambus RIMMs while the P4 uses inexpensive, commodity DDR
SDRAM DIMMs. It is more likely that the Xeon memory will be protected

from errors, as Bill said.

I might have missed one, but I don't believe there were any commercially
available P4 Xeon chipsets that used RDRAM...

/daytripper

Really? I thought intel went 100% RDRAM for Pentium 4 and didn't fix
that gaff until the second generation P4s came out (Northwood). It
seems most unusual that they would go 100% RDRAM on the desktop claiming
it gave faster memory access (at the penalty of greater price most
consumers weren't willing to accept) but not offer RDRAM on the server
(where all that matters is performance and reliability, price be
damned). Was there no ECC version of RDRAM? That would be a logical
reason to skip it on server platforms. I only remember that intel in
the Willamette days was all RDRAM, so I assumed the Willamette-based
Xeons also were tied to this memory.

Alex
--
My words are my own. They represent no other; they belong to no other.
Don't read anything into them or you may be required to compensate me
for violation of copyright. (I do not speak for my employer.)

"Alex Johnson" wrote in message
...

Really? I thought intel went 100% RDRAM for Pentium 4 and didn't fix that
gaff until the second generation P4s came out (Northwood). It seems most
unusual that they would go 100% RDRAM on the desktop claiming it gave
faster memory access (at the penalty of greater price most consumers
weren't willing to accept) but not offer RDRAM on the server (where all
that matters is performance and reliability, price be damned).

The detail you missed is that RDRAM provides faster memory (although at
higher latency) at a higher price only for fairly small amounts of memory
(256Mb or less). So it was quite sensible on desktops at the time. With
SDRAM, you get your speed by having lots of chips/banks. That isn't very
practical when you're building a video game that only needs 32MB or a
desktop that only gets 256Mb.

DS

On Thu, 06 Jan 2005 13:52:31 -0500, Alex Johnson wrote:

daytripper wrote:
On Wed, 05 Jan 2005 08:16:00 -0500, Alex Johnson wrote:


Bill Davidsen wrote:

Bill wrote:


I am trying to decide between a 1.8 gHz Xeon and a 2.8 gHz P4 for a
single CPU file/print/database server for a small office. So far I have
not found any clear explanation of the differences in or advantages of
on over the other.

Can anyone provide any information or references?


The Xeon fits in a server mobo which supports ECC memory. It is
therefore more reliable, both in terms of MTTF and undetected data
errors. That may justify the small extra cost depending on the cost of
errors, particularly undetected ones.

The P4 is probably less expensive and slightly faster.


The P4 will be more than slightly faster, I expect. Faster bus, larger
cache, some of the performance bottlenecks inside were fixed (it is
Northwood or Prescott compared to the Willamette-based Xeon).

The Xeon probably supports two processors, so you can put a second
1.8GHz Xeon in it. It is possible, but unlikely, that the Xeon allows
up to four processors.

The memory types probably differ a great deal. The Xeon probably uses
expensive Rambus RIMMs while the P4 uses inexpensive, commodity DDR
SDRAM DIMMs. It is more likely that the Xeon memory will be protected

from errors, as Bill said.

I might have missed one, but I don't believe there were any commercially
available P4 Xeon chipsets that used RDRAM...

/daytripper

Really? I thought intel went 100% RDRAM for Pentium 4 and didn't fix
that gaff until the second generation P4s came out (Northwood).

"P4" /= "P4 Xeon"

seems most unusual that they would go 100% RDRAM on the desktop claiming
it gave faster memory access (at the penalty of greater price most
consumers weren't willing to accept) but not offer RDRAM on the server
(where all that matters is performance and reliability, price be
damned).

Sure it makes sense: RDRAM was too expensive for what it offered, and the Xeon
server market is price sensitive - 1u pizza boxen need to be cheap to sell.

Was there no ECC version of RDRAM?

Yes, there were ECC versions of RDRAM.

That would be a logical reason to skip it on server platforms. I only remember that intel in
the Willamette days was all RDRAM, so I assumed the Willamette-based
Xeons also were tied to this memory.

As I indicated, I'm not 100% positive there were no P4 Xeon chipsets with
RDRAM, but I'm pretty sure. The pertinent information would be on the Intel
Developer pages, should you like to check further...

/daytripper