Decisions, ABIT or ASUS??

Hi Troops,
I was just about set on the ABIT IC7-Max3 when I discovered (rather late)
that there were no serial or parallel ports and no support for Motherboard
Monitor 5. I suppose I could use USB to parallel/serial adapters for my HP4
Laser printer and my US Robotics external modem but I'd rather not add any
unnecessary complexity to my new system and I'm not to enthusiastic ?Guru.
Now I'm considering the ASUS P4C800E-Deluxe. Is that the ASUS flagship
board? I seems that the ASUS is not quite as picky about memory as the ABIT
and it's faster - true? I've been an ABIT user through my last five MBs and
I was happy with each. I'm a bit reluctant to switch and leave this
magnificent group I plan to OC a P4 2.8 or 3.0C to 3.3 + (nothing
extreme) with a Zalman CNPS 7000A-Cu cooler. Are these boards so close that
it's a "pickem" choice or is one clearly better? Do you think ABIT will
reconsider and allow MBM5? Finally, are the current P4C800E still revision
1 or are the rev2 boards out?
--
Tally Ho!
Ed

"Ed Forsythe" wrote
Hi Troops,
I was just about set on the ABIT IC7-Max3 when I discovered (rather late)
that there were no serial or parallel ports and no support for Motherboard
Monitor 5.


Hi Ed,

where did you hear about the MAX3 not being supported by MBM?. I think it
should be fine. Having said that how come you are opting for the MAX3? do
you need all that extra SATA support? Did you read any reviews on the new
AI7?

On the ASUS side of things, again it's the Springdale board (P4P800 family)
that are getting all the rave reviews. I heard ASUS boards were generally
better for Intel® Pentium® 4 processors, whereas ABIT are reported to be
better on the AMD Athlon side of things. Due to the fact that I run ABIT
boards over the past 4-5 years, I went for the ABIT AI7 (+Strong supporting
cast). . .
--
Wayne ][
Concepts, Theory, Learning Curves, and woman with big bOObs!

Hi Wayne,
I think you're correct about the MBM5 compatibility issue. Since the Max3
doesn't have uGuru MBM should be OK . The AI7 is a great MB but I may try
4-drive SATA RAID eventually. I like the CAS Gigabyte LAN, no uGURU, Secure
IDE, PAT (does the AI7 have it enabled?), OTES, *AND* the black PCB + the
OTES LEDs increase the cool factor significantly VBG. It should look fab
in a Lian-Li PC-75 case with that gigantic side window . The package
bundle ain't too shabby either. I guess I'm just a sucker for marketing and
the "Flagship" label ;-)) CYA -
--
Tally Ho!
Ed
"Wayne Youngman" wrote in message
...

"Ed Forsythe" wrote
Hi Troops,
I was just about set on the ABIT IC7-Max3 when I discovered (rather
late)
that there were no serial or parallel ports and no support for
Motherboard
Monitor 5.


Hi Ed,

where did you hear about the MAX3 not being supported by MBM?. I think it
should be fine. Having said that how come you are opting for the MAX3? do
you need all that extra SATA support? Did you read any reviews on the new
AI7?

On the ASUS side of things, again it's the Springdale board (P4P800
family)
that are getting all the rave reviews. I heard ASUS boards were generally
better for Intel® Pentium® 4 processors, whereas ABIT are reported to be
better on the AMD Athlon side of things. Due to the fact that I run ABIT
boards over the past 4-5 years, I went for the ABIT AI7 (+Strong
supporting
cast). . .
--
Wayne ][
Concepts, Theory, Learning Curves, and woman with big bOObs!

"Ed Forsythe" wrote
Hi Wayne,
I think you're correct about the MBM5 compatibility issue. Since the Max3
doesn't have uGuru MBM should be OK . The AI7 is a great MB but I may
try
4-drive SATA RAID eventually. I like the CAS Gigabyte LAN, no uGURU,
Secure
IDE, PAT (does the AI7 have it enabled?), OTES, *AND* the black PCB + the
OTES LEDs increase the cool factor significantly VBG. It should look fab
in a Lian-Li PC-75 case with that gigantic side window . The package
bundle ain't too shabby either. I guess I'm just a sucker for marketing
and
the "Flagship" label ;-)) CYA -


Hi,
I am led to believe that the whole *PAT* thing is nothing more than
*MARKETING* and in fact both i875P and i865PE are basically the same board.
I don't think PAT exists (I can here the *GASP*!). Anyway I do like the
look of the MAX3 but I wouldn't use the extras it brings. Sure the CSA is
nice but surely would you need your whole LAN to be able to support this
speed (Hubs, switches, other machines etc). The OTES system cools down the
Mosfets and extracts the warm air from the case, hmm ok not bad, but if you
have a mega ventilated case then surely this will do the job as well. Ok ok
you got me on the *Looks Cool* factor, defo with a side-panel :P. Secure
IDE? wtf?????

So I can't say a bad thing about the MAX3 except its not the best choice for
overclocking (IMO) and its not the best value unless you use all the added
features. I will be constructing the AI7 build anyday now, so I will report
back. . . . .
--
Wayne ][
Concepts, Theory, Learning Curves, and woman with big bOObs!

"Wayne Youngman" wrote in message
...

I am led to believe that the whole *PAT* thing is nothing more than
*MARKETING* and in fact both i875P and i865PE are basically the same
board.
I don't think PAT exists (I can here the *GASP*!).

I'm not gasping, but I am grinning. PAT is a tricky subject because Intel
uses smoke and mirrors to hide what they are really doing. But the
performance benefit is real. You CAN tell the difference in memory
bandwidth.

On an 875 PAT is built in and always on. On an 865 PAT is emulated with bios
settings that may or may not work.

Reading the reviews from sites where hardware experts with cherry-picked
boards and memory get amazing results can be quite different from reading
the forums and newsgroups where everyday users struggle with Abit GAT and
Asus HyperPath emulations. Sometimes they work, sometimes they don't. Many
times in overclocked and highly tweaked systems, they don't.

To get the performance of PAT on an IS7, for example, you need to tweak the
GAT settings, but many people find they can't get stable at anything beyond
the all-Auto defaults and it doesn't work at any ratio other than 1:1.

When I was shopping, the IS7 and IC7 were $20 apart. I spent the extra money
for the 875 and was very glad I did. GAT settings severely limit my
overclocking ability, but I can leave them at the defaults and jack my
system sky-high without losing PAT. How much difference does it make? With
PAT disabled at 275 fsb running 5:4, 2,3,2,6, I get Sandra buffered
bandwidth scores of about 5,000. With PAT enabled using the same settings, I
get about 5750, a 15% improvement that is pretty close to what others report
in similar circumstances.

Maybe the new 865 boards are better at PAT emulation than the old ones, but
IMHO no emulation is better and will always be less reliable than the real
thing.

"Skid" wrote
Maybe the new 865 boards are better at PAT emulation than the old ones,
but
IMHO no emulation is better and will always be less reliable than the real
thing.


Hi Skid,
I don't think the word *emulation* is correct. Have a read about PAT here,
and keep an open mind:

http://tinyurl.com/2aw3l

--
Wayne ][
Concepts, Theory, Learning Curves, and woman with big bOObs!

Wayne Youngman wrote:
"Skid" wrote

Maybe the new 865 boards are better at PAT emulation than the old ones,

but

IMHO no emulation is better and will always be less reliable than the real
thing.



Hi Skid,
I don't think the word *emulation* is correct. Have a read about PAT here,
and keep an open mind:

http://tinyurl.com/2aw3l


IMO that article uses as much 'marketing hype' as Intel is accused of.

It's simply a matter of internal chipset timings and it isn't surprising
that you might want different timings, and buffers, for an 800 MHz bus than
a 533, depending on the chipset characteristics.

Intel found out they could use the 'faster' timings, I.E. something similar
to the 533 settings, on the newer chipsets, and gave it a 'name'. Some
folks have apparently figured out they could 'trick' other chipsets into
doing something similar which is, in one sense, 'overclocking'.

Intel tests and guarantees it works on the ones they support, they don't on
the others, so it's not surprising they don't want the 'name' used on the
ones they don't support because it doesn't, technically, meet spec.

It's interesting that in a group which regularly pushes their components to
'out of spec' speeds that doing the same thing with the chipset gives rise
to theories of some 'sinister plot' on Intel's part.

It's not much different than the venerable BX chipset being officially a
"100 MHz FSB" chipset but Abit, as one example, making the BX133 that could
run 133 MHz.

I've also observed the 'debate' on what the 'official Intel specs' are for
how much one can 'overclock' with it still 'enabled'. There IS no such
thing because there is no 'official' overclocking. Intel expects the thing
to be run as specified, in all regards, and anything out of spec is
whatever some third party manufacturer decides to implement on their own
and what they allow in the BIOS version they provide.

"Wayne Youngman" wrote in message
...

"Skid" wrote
Maybe the new 865 boards are better at PAT emulation than the old ones,
but
IMHO no emulation is better and will always be less reliable than the
real
thing.


Hi Skid,
I don't think the word *emulation* is correct. Have a read about PAT
here,
and keep an open mind:

http://tinyurl.com/2aw3l

I've seen that one, and basically all it shows is that at 200 fsb a good
Springdale using PAT emulation can keep up with a good Canterwood. Toss in
overclocking and ratios, and all bets are off. Notice that the highest
scores in those benchmarks are from Springdales running at 300 fsb, 3:2,
with PAT emulation disabled.

The Canterwood boards were not tested at speeds higher than 250, for reasons
the reviewer doesn't explain. My Canterwood does 275 at 5:4 before the cpu
gives out, and my Sandra memory scores are within an eyelash of the
Springdale at 300. Still apples to oranges.

Believe me, I have kept an open mind on the issue. I read all the same
reviews when I was making my choice almost a year ago. Things are as
confusing now as then, with one exception. Reading posts from end-users in
the forums shows a lot of people can't get the results reviewers report on
tweaked Springdales, especially when overclocking.

There's a slightly more technical explanation of what PAT is and how Asus
and Abit emulate it at http://www.lostcircuits.com/motherboard/asus_p4p800/

After praising Asus for fine-tuning memory performance on the Springdale,
the reviewer offers this caution:
"Of course, there is a problem with this scheme, usually wafers are tested
to make sure that the speed bin will be able to withstand the high operating
frequency at reduced pipeline stages. Those wafers that pass are binned as
Canterwood and packaged using the more elaborate 1005 BGA package. Keep in
mind though that only a small fraction of die is tested, and that the
qualification means passing speed tests way in excess of the nominal
frequency.

This leaves a waste majority of chips that will pass at the nominal value
but with small margins and probably more that just fall through the cracks
because they were not tested at all. What it all comes down to is the fact
that most Springdale PE chipsets will more or less meet the PAT
requirements."

IMHO, it still comes down to a sure thing versus a solid maybe.

Skid wrote:
"Wayne Youngman" wrote in message
...

"Skid" wrote

Maybe the new 865 boards are better at PAT emulation than the old
ones,

but

IMHO no emulation is better and will always be less reliable than
the

real

thing.


Hi Skid, I don't think the word *emulation* is correct. Have a read
about PAT

here,

and keep an open mind:

http://tinyurl.com/2aw3l


I've seen that one, and basically all it shows is that at 200 fsb a good
Springdale using PAT emulation can keep up with a good Canterwood. Toss
in overclocking and ratios, and all bets are off. Notice that the
highest scores in those benchmarks are from Springdales running at 300
fsb, 3:2, with PAT emulation disabled.

The Canterwood boards were not tested at speeds higher than 250, for
reasons the reviewer doesn't explain. My Canterwood does 275 at 5:4
before the cpu gives out, and my Sandra memory scores are within an
eyelash of the Springdale at 300. Still apples to oranges.

Believe me, I have kept an open mind on the issue. I read all the same
reviews when I was making my choice almost a year ago. Things are as
confusing now as then, with one exception. Reading posts from end-users
in the forums shows a lot of people can't get the results reviewers
report on tweaked Springdales, especially when overclocking.

There's a slightly more technical explanation of what PAT is and how
Asus and Abit emulate it at
http://www.lostcircuits.com/motherboard/asus_p4p800/

After praising Asus for fine-tuning memory performance on the
Springdale, the reviewer offers this caution: "Of course, there is a
problem with this scheme, usually wafers are tested to make sure that
the speed bin will be able to withstand the high operating frequency at
reduced pipeline stages. Those wafers that pass are binned as Canterwood
and packaged using the more elaborate 1005 BGA package. Keep in mind
though that only a small fraction of die is tested, and that the
qualification means passing speed tests way in excess of the nominal
frequency.

This leaves a waste majority of chips that will pass at the nominal
value but with small margins and probably more that just fall through
the cracks because they were not tested at all. What it all comes down
to is the fact that most Springdale PE chipsets will more or less meet
the PAT requirements."

IMHO, it still comes down to a sure thing versus a solid maybe.

Which also gets to why Intel doesn't enable PAT on the other chipsets. I
mean, how would you feel about buying a chipset with the 'spec' of "it
probably works, maybe" (too bad about the one you got)?

That was one of the best articles I've seen on the subject except for, IMO,
a characterization of 'testing' and 'margins' that tends to create a
potentially false illusion. As alluded to in the 'maybe spec', Intel has to
test them for the complete range of operation including the 'margins' of
numerous parameters, many of which are external such as motherboard design,
electrical margins, environment, etc. Restrict your system to operating at
a lower temperature (than the full spec range) and you might be able to
shift some of that 'margin' to the chipset speed (as we do when
overclocking processors). If the motherboard manufacturer's PCB design is
on the 'better' end of the tolerances (capacitances, propagation delays,
etc.) then you might be able to shift some of that 'margin' to the chipset
speed. If your power rails are held to a tighter spec (stability, noise,
surge, droop, etc.) then you might be able to shift some of that 'margin'
to the chipset speed, and so on. But Intel has to guarantee proper
operation over the entire range of all of them, even when they all conspire
to the 'worst case'. That means if you look at just the 'average', or
'typical', conditions it may seem like 'over testing' and an 'excessive
margin' but that impression is misleading for the reasons just stated.