Greetings to all,

Has anyone heard of any networked desktop problems that have been proven to
be due to AMD processors?

I'm a network admin, and I've got a consultant who is advising that we
purchase some AMD powered desktops. In the past my attitude has been to
always purchase pentiums - simply because I do NOT have the time to worry
about possible processor conflicts (there's FAR too many other issues for
admins to fret about!) Not that pentiums are perfect, but the user base is
so massive that it's likely enormous pressure will be brought to bear on
those responsible for any problems. It's also my hunch that any performance
gain from AMD would not be significant to my users. Finally, a pentium
problem is much easier to explain to management than an AMD problem.

I should mention that we are using Win2k servers, Win2k3 servers, Citrix
(bigtime), AD, VPNs across several sites, wireless networking, SQL, and
(soon) considerable multimedia technology.


Thanks!

Using a finger dipped in purple ink, "Frankster" <(E-Mail Removed)> scribed:

>Greetings to all,
>
>Has anyone heard of any networked desktop problems that have been proven to
>be due to AMD processors?

I have 7 home systems networked.

5 are AMD-based, 2 are Intel.

NO networking problems that can't be blamed on MS,
and few even of those.

You're FAR more likely to find a problem with software
than you are with a CPU.








--

The truth is out there,

but it's not interesting enough for most people.

On Tue, 3 May 2005 17:22:27 -0400, "Frankster" <(E-Mail Removed)> wrote:

>Greetings to all,
>
>Has anyone heard of any networked desktop problems that have been proven to
>be due to AMD processors?
>
>I'm a network admin, and I've got a consultant who is advising that we
>purchase some AMD powered desktops. In the past my attitude has been to
>always purchase pentiums - simply because I do NOT have the time to worry
>about possible processor conflicts (there's FAR too many other issues for
>admins to fret about!) Not that pentiums are perfect, but the user base is
>so massive that it's likely enormous pressure will be brought to bear on
>those responsible for any problems. It's also my hunch that any performance
>gain from AMD would not be significant to my users. Finally, a pentium
>problem is much easier to explain to management than an AMD problem.
>
>I should mention that we are using Win2k servers, Win2k3 servers, Citrix
>(bigtime), AD, VPNs across several sites, wireless networking, SQL, and
>(soon) considerable multimedia technology.
>
>
>Thanks!
>
Never ever heard of any network problem traceable to a CPU, be it AMD,
Intel, or anything else. OTOH, had quite a few of them due to NICs,
hubs, and other networking gear malfunctioning, including, but not
limited to, Intel-branded hardware.

>
> I should mention that we are using Win2k servers, Win2k3 servers, Citrix
> (bigtime), AD, VPNs across several sites, wireless networking, SQL, and
> (soon) considerable multimedia technology.
>
>


My understanding is the amd64 cpu is an extension of the original athlon
design and as such has a longer track record than the pentium4, that said
i fail to see how a cpu would have any effect on networking except speed
{1} you should be more concerned with the chipsets of the motherboards
though even then i doubt you would have problems.

even if you have apps compiled for only pentium 4's {2} an amd64 has sse2
and some the major bits of sse3 as well though i highly doubt you have any
code that tweaked

the reason the consultant is recomending amd is the bitch slap intel in
every real test except transcoding so badly that its not even funny. for
server there hasnt been a reason to chose intel for 18months and it will
be next year before their server chips go dual core, its unfortunate since
servers are one of the places dual cores really thrive.

you may wish to expand upon the multimedia stuff you will be doing because
its one of the few areas intel can still compete.


{1} i suppose if there was a major bug and the os was coded to have probs
with that particular bug then possibly it could give you probs but there
would be a patch to solve it like the pentium floating point thing


{2} -march=pentium4 or whatever icc uses to force the use of the sse2 unit

Frankster wrote:
> Greetings to all,
>
> Has anyone heard of any networked desktop problems that have been proven to
> be due to AMD processors?

What possible network problem could ever come from the CPU? Your
question makes about as much sense as asking whether having too many
Toshiba rather than Sony tv's tuned to a certain channel has ever caused
that tv station to go offline.


> I'm a network admin, and I've got a consultant who is advising that we
> purchase some AMD powered desktops. In the past my attitude has been to
> always purchase pentiums - simply because I do NOT have the time to worry
> about possible processor conflicts (there's FAR too many other issues for
> admins to fret about!) Not that pentiums are perfect, but the user base is
> so massive that it's likely enormous pressure will be brought to bear on
> those responsible for any problems. It's also my hunch that any performance
> gain from AMD would not be significant to my users. Finally, a pentium
> problem is much easier to explain to management than an AMD problem.

It sounds to me like you want us to confirm your prejudices. You don't
need us to do that, all you have to do is stomp into your management's
office and tell them, "I hate AMD's, I ain't gonna work in no office
that has them dang AMD's." I'm sure they'll find that to be a compelling
argument. :-)

But if you are open-minded, then there are some very tangible reasons
for going with AMD processors instead of Intel ones. Lower power
consumption, resulting in less electrical bills, and less heat. Recent
tests measured right at the power supply have shown that with
identically configured AMD and Intel systems, the AMD systems would use
185W while the Intel ones would use 250W. This is a result of not only
the greater power requirements of the CPU, but the extra fans and longer
durations that the fans must be run at full speed on the Intel systems.

As for performance, and whether users will notice the difference, that
entirely depends on what the current systems that they are comparing
them to. They will easily notice a difference if the current systems are
P4 2.4Ghz or lower. Upto P4 3.0Ghz, they might notice a slight
difference. Beyond that, I doubt it, especially for office apps. One
thing they may find is that the AMD's are much more balanced performers,
they are usually rated "very good" at all applications. Whereas the
Pentium 4's maybe merely "good" at most applications, and "excellent" in
a few select applications like video encoding.

>
> I should mention that we are using Win2k servers, Win2k3 servers, Citrix
> (bigtime), AD, VPNs across several sites, wireless networking, SQL, and
> (soon) considerable multimedia technology.

Excellent, AMD processors are the preferred choice for most server
applications too, especially all of those ones that you just mentioned.
For example, it's well known that Pentium 4/Xeon needs highly optimized
software to run well; an older operating system like Win2K wouldn't be
able to provide that kind of optimization on a Xeon since Microsoft
doesn't update it anymore. An AMD Opteron would take any operating
system, old or new, run it well with or without optimization; because of
its more generally balanced performance. Also Microsoft itself initially
developed the new 64-bit versions of its operating systems entirely on
AMD processors.

Yousuf Khan

On Tue, 3 May 2005 17:22:27 -0400, "Frankster" <(E-Mail Removed)> wrote:

>Greetings to all,
>
>Has anyone heard of any networked desktop problems that have been proven to
>be due to AMD processors?

There would certainly be nothing due to the processors themselves. They
perform flawlessly. The only possible problem you might have would be with
older versions of anti-virus software, e.g. Symantec Anti-Virus 8.01 in
combination with WinXP SP2 - the NX bit does work and you can get DEP (Data
Execute Protection) crashes under some situations. I normally turn off
Floppy Check on Shutdown anyway but if you don't.... BOOM! I am sick to
death of Symantec with their flakey drivers and gouging on "Gold Insurance"
to get "upgrades" which are only bug fixes.

>I'm a network admin, and I've got a consultant who is advising that we
>purchase some AMD powered desktops. In the past my attitude has been to
>always purchase pentiums - simply because I do NOT have the time to worry
>about possible processor conflicts (there's FAR too many other issues for
>admins to fret about!) Not that pentiums are perfect, but the user base is
>so massive that it's likely enormous pressure will be brought to bear on
>those responsible for any problems. It's also my hunch that any performance
>gain from AMD would not be significant to my users. Finally, a pentium
>problem is much easier to explain to management than an AMD problem.

Huh? Is your "management" just too dense to understand "equipment"? You
can get a dud from any mfr.

>I should mention that we are using Win2k servers, Win2k3 servers, Citrix
>(bigtime), AD, VPNs across several sites, wireless networking, SQL, and
>(soon) considerable multimedia technology.

We have several AMD systems, the most recent K8s based on mbrds with VIA
K8T800 chipsets and nVidia nForce3 chipsets with integrated Gigabit NICs;
we also have some K7s on nForce2 chipsets with integrated 100Mbit NICs.
Not a problem with any of them. I recently upgraded our LAN hub stack to a
D-Link DGS-1224TG 10/100/1000Mbps switch and all the integrated 100Mbps
NICs are working fine at 100Mbps full duplex and the Gigabit NICs at
1000Mbps full duplex... all on our old cat-5 cabling. The port monitoing
on the switch is showing very low error rates on all ports - I figure what
errors there are, are probably due to power-on/off resyncing.

The old file/print server is about to be upgraded to a K8 on a nForce4
chipset... to keep up with the PCs.:-)

--
Rgds, George Macdonald

Frankster wrote:

>Has anyone heard of any networked desktop problems that have been proven to
>be due to AMD processors?
>
>I'm a network admin,

A network admin as clueless as this? Tell me you're joking.

On Wed, 04 May 2005 08:37:09 -0500, chrisv <(E-Mail Removed)> wrote:

>Frankster wrote:
>
>>Has anyone heard of any networked desktop problems that have been proven to
>>be due to AMD processors?
>>
>>I'm a network admin,
>
>A network admin as clueless as this? Tell me you're joking.

I blame Microsoft......and "wizards" ;-)

On Wed, 04 May 2005 09:46:51 -0400, daytripper wrote:

> On Wed, 04 May 2005 08:37:09 -0500, chrisv <(E-Mail Removed)> wrote:
>
>>Frankster wrote:
>>
>>>Has anyone heard of any networked desktop problems that have been proven to
>>>be due to AMD processors?
>>>
>>>I'm a network admin,
>>
>>A network admin as clueless as this? Tell me you're joking.
>
> I blame Microsoft......and "wizards" ;-)

Could be that nine year old Microsoft Certified Professional!

http://www.pcpro.co.uk/news/72339/ni...fessional.html

All this time I thought network and Microsoft and Desktop was a no no in
one sentence. Especially after reading this:

http://www.zdnet.com.au/news/securit...9189876,00.htm

This is like the never ending online running joke of the decade.
Regardless if they would of only listened.

Gnu_Raiz

On Wed, 04 May 2005 09:46:51 -0400, daytripper <(E-Mail Removed)>
wrote:

>On Wed, 04 May 2005 08:37:09 -0500, chrisv <(E-Mail Removed)> wrote:
>
>>Frankster wrote:
>>
>>>Has anyone heard of any networked desktop problems that have been proven to
>>>be due to AMD processors?
>>>
>>>I'm a network admin,
>>
>>A network admin as clueless as this? Tell me you're joking.
>
>I blame Microsoft......and "wizards" ;-)

What? You mean a network admin who's never had to drop down to all the
command line, character mode err, utilities?

--
Rgds, George Macdonald