|
|
|
|
| Author |
Message |
Carlos Moreno
Guest
|
 Posted: Mon Jul 14, 2003 9:31 pm Post subject: Athlon MP -- really a solid option? |
 |
|
Hi,
I'm about to decide to upgrade my machine (currently have an
Athlon 1GHz), and the main option in my mind is a dual Athlon,
but there is some confusion and doubts about it.
AMD's FAQ's talk about considerable advantages for the 2800+
model (Barton). As I seem to understand, the FSB for the
Barton chips is 333MHz; however, the available motherboards
(including my main candidate, Tyan Tiger MPX) only support
200/266, and PC1600/PC2100 memory. Shouldn't one be able to
use PC2700 running a 333MHz FSB?
Overall, is it worth the extra bucks for the 2800+ over the,
say, 2400+ ? (the 2400+ is a very beautiful number -- internal
clock at exactly 2GHz... Seems a nice choice, and the price is
about half the price of the 2800+ (and thus, it becomes four
times, since I need to buy two processors (*)) but I wonder
if the advantages of the 2800+ are as solid as to justify the
higher price)
More in general -- and I know what I'm about to say may be
extremely silly, mainly coming from a huge AMD advocate and
Intel "anti-advocate" as I am -- how does a dual Athlon
system compares to a dual Xeon system? I have always
assumed that it's a non-contest that AMD would win hands
down... But recent comments and articles I've been reading
kind of wake up some doubts in my mind... Is it a good
choice?
BTW, I'm going to run Linux (mainly), and Windows 2000 for
some video editing and DVD recording.
Thanks for any comments,
Carlos
--
(*) Yes, that was a joke! |
|
| Back to top |
|
 |
|
|
JK
Guest
|
| Posted: Mon Jul 14, 2003 10:18 pm Post subject: Re: Athlon MP -- really a solid option? |
|
|
Carlos Moreno wrote:
| Quote: | Hi,
I'm about to decide to upgrade my machine (currently have an
Athlon 1GHz), and the main option in my mind is a dual Athlon,
but there is some confusion and doubts about it.
AMD's FAQ's talk about considerable advantages for the 2800+
model (Barton). As I seem to understand, the FSB for the
Barton chips is 333MHz; however, the available motherboards
(including my main candidate, Tyan Tiger MPX) only support
200/266, and PC1600/PC2100 memory. Shouldn't one be able to
use PC2700 running a 333MHz FSB?
Overall, is it worth the extra bucks for the 2800+ over the,
say, 2400+ ? (the 2400+ is a very beautiful number -- internal
clock at exactly 2GHz... Seems a nice choice, and the price is
about half the price of the 2800+ (and thus, it becomes four
times, since I need to buy two processors (*)) but I wonder
if the advantages of the 2800+ are as solid as to justify the
higher price)
|
Why not go for the MP2600+? It is priced much closer to
the MP2400+ than the MP2800+.
| Quote: |
More in general -- and I know what I'm about to say may be
extremely silly, mainly coming from a huge AMD advocate and
Intel "anti-advocate" as I am -- how does a dual Athlon
system compares to a dual Xeon system?
|
Doing what? if your budget is not so small, you might want
to consider a dual Opteron system instead?
| Quote: | I have always
assumed that it's a non-contest that AMD would win hands
down... But recent comments and articles I've been reading
kind of wake up some doubts in my mind... Is it a good
choice?
|
What are you doing? For some applications, you might be
better off on a cost benefit basis by using a single XP3000+ 400
instead?
| Quote: |
BTW, I'm going to run Linux (mainly), and Windows 2000 for
some video editing and DVD recording.
|
You might want to consider a dual Opteron or even one way
Opteron system. I haven't seen any motherboards for the Opteron 140
yet, however I guess they might be available soon? The dual Opteron
motherboards look like they start at around $450.
| Quote: |
Thanks for any comments,
Carlos
--
(*) Yes, that was a joke! |
|
|
| Back to top |
|
|
Stacey
Guest
|
| Posted: Tue Jul 15, 2003 5:22 am Post subject: Re: Athlon MP -- really a solid option? |
|
|
JK wrote:
| Quote: |
Carlos Moreno wrote:
More in general -- and I know what I'm about to say may be
extremely silly, mainly coming from a huge AMD advocate and
Intel "anti-advocate" as I am -- how does a dual Athlon
system compares to a dual Xeon system?
Doing what? if your budget is not so small, you might want
to consider a dual Opteron system instead?
|
LOL you're really going to get an objective opinion from JK!
--
Stacey |
|
| Back to top |
|
|
Rob Stow
Guest
|
| Posted: Tue Jul 15, 2003 1:19 pm Post subject: Re: Athlon MP -- really a solid option? |
|
|
Stacey wrote:
| Quote: | JK wrote:
Carlos Moreno wrote:
More in general -- and I know what I'm about to say may be
extremely silly, mainly coming from a huge AMD advocate and
Intel "anti-advocate" as I am -- how does a dual Athlon
system compares to a dual Xeon system?
Doing what? if your budget is not so small, you might want
to consider a dual Opteron system instead?
LOL you're really going to get an objective opinion from JK!
|
It might not be an "objective opinion", but as it pertains to
this question posed by the OP, JK certainly asked an obvious
and relevant question. If there is room in the OP's budget,
an Opteron dualie definitely deserves consideration.
The way you dismissed the Opteron question merely because it
came from JK instead of considering the question on its own
merits suggests that when it comes to JK you are lacking in
objectivity :-D |
|
| Back to top |
|
|
The little lost angel
Guest
|
| Posted: Wed Jul 16, 2003 12:17 am Post subject: Re: Athlon MP -- really a solid option? |
|
|
On Mon, 14 Jul 2003 20:22:20 -0400, Stacey <fotocord@yahoo.com> wrote:
| Quote: | LOL you're really going to get an objective opinion from JK!
|
Well, Stacey, this time round, JK hasn't actually played the advocate
yet. Since the OP was asking about AMD based systems in general, the
Opteron suggest was pretty on the mark.
I was thinking of asking the OP to check out the pricing of Opterons
vs the MP too!
--
L.Angel: I'm looking for web design work.
If you need basic to med complexity webpages at affordable rates, email me 
Standard HTML, SHTML, MySQL + PHP or Javascript. No ASP, CF or Frontpage. |
|
| Back to top |
|
|
Fix your Windows Problems - FAST.
FREE Safe Scan Registry Check. Locate & Fix Errors in Minutes!
|
|
Stacey
Guest
|
| Posted: Thu Jul 17, 2003 7:27 am Post subject: Re: Athlon MP -- really a solid option? |
|
|
The little lost angel wrote:
| Quote: | On Mon, 14 Jul 2003 20:22:20 -0400, Stacey <fotocord@yahoo.com> wrote:
LOL you're really going to get an objective opinion from JK!
Well, Stacey, this time round, JK hasn't actually played the advocate
yet. Since the OP was asking about AMD based systems in general, the
Opteron suggest was pretty on the mark.
|
Jk's coment was in responce to the OP's question:
"More in general -- and I know what I'm about to say may be extremely silly,
mainly coming from a huge AMD advocate and Intel "anti-advocate" as I am --
how does a dual Athlon system compares to a dual Xeon system?"
Just thought it was funny he never answered the question, just
sugested/pushed another AMD product.
--
Stacey |
|
| Back to top |
|
|
Stacey
Guest
|
| Posted: Thu Jul 17, 2003 7:30 am Post subject: Re: Athlon MP -- really a solid option? |
|
|
Rob Stow wrote:
| Quote: |
The way you dismissed the Opteron question merely because it
came from JK instead of considering the question on its own
merits suggests that when it comes to JK you are lacking in
objectivity :-D
|
Guilty as charged!
Given the OP asked "how does a dual Athlon system compares to a dual Xeon
system?" and JK once again ignored the question to push AMD products just
seemed comical...
--
Stacey |
|
| Back to top |
|
|
Tony Hill
Guest
|
| Posted: Sun Jul 20, 2003 2:25 am Post subject: Re: Athlon MP -- really a solid option? |
|
|
On Mon, 14 Jul 2003 12:31:45 -0400, Carlos Moreno
<moreno_at_mochima_dot_com@x.xxx> wrote:
| Quote: | I'm about to decide to upgrade my machine (currently have an
Athlon 1GHz), and the main option in my mind is a dual Athlon,
but there is some confusion and doubts about it.
AMD's FAQ's talk about considerable advantages for the 2800+
model (Barton). As I seem to understand, the FSB for the
Barton chips is 333MHz; however, the available motherboards
(including my main candidate, Tyan Tiger MPX) only support
200/266, and PC1600/PC2100 memory. Shouldn't one be able to
use PC2700 running a 333MHz FSB?
|
The AthlonXP "Barton" chips run at 333MHz or 400MHz bus speeds (err,
166 and 200MHz DDR bus speeds respectively). The AthlonMP "Barton"
chips run at a 133/266MHz DDR bus speed. As such, the only difference
between the "Thoroughbred" and "Barton" AthlonMP chips is the amount
of L2 cache. The older "Thoroughbred" models (all of those up to the
AthlonMP 2600+) have 256KB of L2 cache, while the "Bartons" (the new
AthlonMP 2800+ and any possible future chips) have 512KB of L2.
In single processor desktop systems, the extra L2 cache tended to
added about 4 or 5% to the performance on average, though some
applications saw much larger or smaller differences. In a
multi-processor setup, my guess is that you'll see a slightly larger
increase in performance from the extra cache (probably 5-6% on
average) since memory bus contention is higher in dual-processor
systems.
| Quote: | Overall, is it worth the extra bucks for the 2800+ over the,
say, 2400+ ? (the 2400+ is a very beautiful number -- internal
clock at exactly 2GHz... Seems a nice choice, and the price is
about half the price of the 2800+ (and thus, it becomes four
times, since I need to buy two processors (*)) but I wonder
if the advantages of the 2800+ are as solid as to justify the
higher price)
More in general -- and I know what I'm about to say may be
extremely silly, mainly coming from a huge AMD advocate and
Intel "anti-advocate" as I am -- how does a dual Athlon
system compares to a dual Xeon system? I have always
assumed that it's a non-contest that AMD would win hands
down... But recent comments and articles I've been reading
kind of wake up some doubts in my mind... Is it a good
choice?
BTW, I'm going to run Linux (mainly), and Windows 2000 for
some video editing and DVD recording.
|
Well, the Xeon solutions are now much more attractive than they once
were since the price has dropped considerably on their motherboards.
Where you used to pay about $250 for a dual AthlonMP board vs. $550+
for a dual Xeon board, now the prices are more like $225 vs. $275
respectively.
For straight DVD recording, both chips are pretty darn quick and can
do the recording part in real-time, you're real bottleneck here is
definitely going to be the hard drive. However, if you're planning on
doing some encoding to a newer, more processor intensive codec than
the MPEG-2 used for DVDs, than that's another story altogether.
Encoding video streams is a VERY processor intensive task, one that
can gobble up tons of processor cycles. Unfortunately it's also a
task that can vary widely from one codec to another as to how it
performs on different processors. There are some codecs that run MUCH
faster on P4's and Xeons than they do on Athlons, and there are others
that run quicker on the AMD setups. Ideally you should try to find
some benchmarks out there that use the software that YOU use to do as
close to what you're planning on doing and see how they compare.
Now, that being said, as a general rule, the Xeons do tend to be a bit
quicker when comparing comperable chips (ie AthlonMP 2400+ vs. Xeon
2.4GHz), however not by nearly as much as when comparing P4's vs.
AthlonXPs. Those who poo-pooed AMD's "Quantispeed" rating system
obviously never tried to compare single vs. dual-processor setups for
AMD vs. Intel. An AthlonXP and a single AthlonMP with the same
quantispeed rating will perform virtually identically. On the other
hand, a 3.0GHz P4 will be quite a bit faster than a 3.0GHz Xeon, since
the Xeon has only a 400MHz or 533MHz bus speed while the P4 have an
800MHz bus speed. Now, why is this important? Well simply because
you're much more likely to find benchmarks comparing a single
AthlonXP to a single P4 than you are to finding a comparison between
dual AthlonMPs vs. dual Xeons.
Now, after all my long-windedness.. If I were building this system
for myself to do the sort of thing you're looking at, I would probably
go for a new dual-Xeon board based off of Intel's new e7505 chipset
(~$300) and get a pair of 2.4GHz, 533MHz bus speed Xeons (~$250 a
piece). This is more expensive than a dual-Athlon setup, where you
could probably get the board for $225 and the AthlonMP 2400+ for about
$175 a piece, but the total price difference once you factor
everything into the whole system cost isn't that much. As I mentioned
above, Xeons will tend to be a bit faster for what you're after than
AthlonMPs at comperable speeds, particularly when dealing with the
533MHz bus speed Xeons (which are fairly new, most of the Xeons out
there are still using 400MHz bus speeds). Also, Intel's new e7505
chipset looks quite solid and has a nice modern feature set to it,
much better than the slightly long-in-tooth AMD 760MPX chipset. One
word of warning about this chipset though, most boards come with only
one or two regular PCI slots, with the rest being PCI-X slots. Most
high-end add-in cards are now available as PCI-X cards, but if you
have a lot of existing PCI cards, this could be a problem. |
|
| Back to top |
|
|
Yousuf Khan
Guest
|
| Posted: Sun Jul 20, 2003 7:44 pm Post subject: Re: Athlon MP -- really a solid option? |
|
|
"Tony Hill" <hilla@uoguelph.ca> wrote in message
news:k2bjhvormnasflch5kkmuv866jbvrr7rbl@4ax.com...
| Quote: | In single processor desktop systems, the extra L2 cache tended to
added about 4 or 5% to the performance on average, though some
applications saw much larger or smaller differences. In a
multi-processor setup, my guess is that you'll see a slightly larger
increase in performance from the extra cache (probably 5-6% on
average) since memory bus contention is higher in dual-processor
systems.
|
Would that really be much of an issue on the Athlons? They do have a
point-to-point bus for dual processors.
Yousuf Khan |
|
| Back to top |
|
|
Tony Hill
Guest
|
| Posted: Mon Jul 21, 2003 3:41 am Post subject: Re: Athlon MP -- really a solid option? |
|
|
On Sun, 20 Jul 2003 14:44:55 GMT, "Yousuf Khan"
<bbbl67.spam@yahoo.com.nospam> wrote:
| Quote: | "Tony Hill" <hilla@uoguelph.ca> wrote in message
news:k2bjhvormnasflch5kkmuv866jbvrr7rbl@4ax.com...
In single processor desktop systems, the extra L2 cache tended to
added about 4 or 5% to the performance on average, though some
applications saw much larger or smaller differences. In a
multi-processor setup, my guess is that you'll see a slightly larger
increase in performance from the extra cache (probably 5-6% on
average) since memory bus contention is higher in dual-processor
systems.
Would that really be much of an issue on the Athlons? They do have a
point-to-point bus for dual processors.
|
For processors yes, but not for memory. Both AthlonMP chips in an AMD
760MPX based dual-processor system share a single, 64-bit wide, PC2100
DDR memory bus. As such, I would tend to guess that the extra cache
would help more in dual-processor AthlonMP systems than single
processor AthlonXP systems, though testing that hypothesis might not
be the easiest task around. |
|
| Back to top |
|
|
Fix your Windows Problems - FAST.
FREE Safe Scan Registry Check. Locate & Fix Errors in Minutes!
|
|
Yousuf Khan
Guest
|
| Posted: Mon Jul 21, 2003 5:26 pm Post subject: Re: Athlon MP -- really a solid option? |
|
|
"Tony Hill" <hilla@uoguelph.ca> wrote in message
news:mm5mhv8tiumk38it1bq05m4gfsdsb7u17d@4ax.com...
| Quote: | Would that really be much of an issue on the Athlons? They do have a
point-to-point bus for dual processors.
For processors yes, but not for memory. Both AthlonMP chips in an AMD
760MPX based dual-processor system share a single, 64-bit wide, PC2100
DDR memory bus. As such, I would tend to guess that the extra cache
would help more in dual-processor AthlonMP systems than single
processor AthlonXP systems, though testing that hypothesis might not
be the easiest task around.
|
Yes, but the memory is controlled by single chipset, and the Athlons will
talk to the chipset in a turn-based point-to-point manner. So the chaotic,
colliding bus contention issue of a traditional Intel chipset is avoided.
Yousuf Khan |
|
| Back to top |
|
|
Scott Alfter
Guest
|
| Posted: Mon Jul 21, 2003 10:05 pm Post subject: Re: Athlon MP -- really a solid option? |
|
|
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
In article <HVxSa.10425$zwL.6086@news04.bloor.is.net.cable.rogers.com>,
Yousuf Khan <bbbl67.spam@yahoo.com.nospam> wrote:
| Quote: | "Tony Hill" <hilla@uoguelph.ca> wrote in message
news:k2bjhvormnasflch5kkmuv866jbvrr7rbl@4ax.com...
In single processor desktop systems, the extra L2 cache tended to
added about 4 or 5% to the performance on average, though some
applications saw much larger or smaller differences. In a
multi-processor setup, my guess is that you'll see a slightly larger
increase in performance from the extra cache (probably 5-6% on
average) since memory bus contention is higher in dual-processor
systems.
Would that really be much of an issue on the Athlons? They do have a
point-to-point bus for dual processors.
|
You still have just one chunk of memory that's shared between both
processors. The Opteron provides for giving each processor its own memory
(by moving the memory controller into the processor), but in an Athlon MP,
all memory requests go through the northbridge. The point-to-point bus can
allow one processor to do an I/O request while the other does a memory
request (something Intel's processors, with their shared bus, can't do), but
if they both want to access memory, one of them will have to wait.
_/_ Scott Alfter
/ v \ salfter@salfter.dyndns.org
(IIGS( http://alfter.us Top-posting!
\_^_/ pkill -9 /bin/laden >What is the most annoying thing on Usenet?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (Linux)
Comment: For info see http://www.gnupg.org
iD8DBQE/HB12VgTKos01OwkRAoixAJ9joy1384SlfWg7lZ3CyVyldUG6iACfd/5B
Gy5YuZ1nm3lNTHaoKyD+tfc=
=hby9
-----END PGP SIGNATURE----- |
|
| Back to top |
|
|
|
  |
Index
-> CPU |
All times are GMT
|
| Page 1 of 1 |
|
| |
Watched Topics
Search
Register
Log in to check your private messages
Profile
Log in
Memberlist
Usergroups