*** Official ASUS P4P800/Deluxe (865PE) Thread ***

[donfm]

The recommended memory list from Asus contains only 2 brands that I see readily available in the USA for a reasonable price, from mainstream sources such as Newegg and Googlegear. The brands are Samsung and Apacer (which is made by Samsung). These will work in all 3 configurations of the memroy slots,... ie. Single channel one slot filled, double channel 2 slots filled, and double channel all 4 memory slots filled according to tests done by Asus.

Note that Corsair PC3200 is NOT on the list and the PC3500 must be in a 512MB stick and is not listed as working in a 4 slot filled double channel memory configuration on the chart. Newegg carries the Apacer at $74 a stick I believe but not sure if that was 512 or 256MB stick. I just returned my Corsair TWINX512-3200LLPT to avoid problems now and in the future. It really sucks for what you pay for Corsair PC3200 and it won't work in a PC3200 board. Like I said previously, if the Corsair rep doesn't recommend Corsair PC3200 memory do you have to be hit over the head to get the picture?? It doesn't work well on this board! Good luck all.

 

[LegionX]

hi again

going to order this board hopefully in the next few hours and with all the talks of memory issues with certain brands/speeds i want to make sure that my 2 sticks of 333 DDR 512mb willl run ok, i wont be buying new memory for awhile - til i can afford both the ddr400+ and the newer 800/3 gig+ p4's.

so will i have any problems with the basic samsung 333's?

also off topic but need to know... i am planning on also buying a celeron 1.7 to use with my old board (msi 648 max-L) and want to make sure it can use a celeron?

any help would be greatly appreciated

 

[morkman100]

Sorry about the mixup. I confused his reply to a post about someone already having the CMX512-3200C2 memory with his memory recommendations for Springdale chipsets.

(In case anybody was interested, RamGuy said that for anyone having problems with the CMX512-3200C2 in a Springdale motherboard, try setting the DDR voltage to 2.75 V and the memory timing to 2-4-4-8).

Thanks for pointing out my mistake.

morkman

 

[paulj2]

No. The "(1.5 V only)" is describing the slot. It has nothing to do with AGP 4X or 8X.

 

[ccgr]

http://www.tomshardware.com/motherboard/20030603/index.html

Asus press and retail versions are different....good article!

 

[donfm]

Originally posted by: ccgr
http://www.tomshardware.com/motherboard/20030603/index.html

Asus press and retail versions are different....good article!

Boy everyone should read that article. It is what I suspected with my MSI NEO-LSR board all along. It is NOT the same board that they sell on the street. Now I feel even better about returning my MSI board. Thanks!

 

[Zarick]

So the question now is... if ASUS is sending different boards to reviews are we not getting a good review.
Are we getting lower quality boards with less performance.

 

[monadian]

Oops. This was meant to be in response to Styler2K's question about the P4P800 and P4P800 Deluxe using the same 3Com Gigabyte LAN chips...

And the answer is:

That's correct.
http://www.asus.com/products/mb/socket478/p4p800/overview.htm

 

[monadian]

I've read through darn near all the posts in this thread and didn't find an answer to one question in particular.

Regarding the differences between the P4P800 versus the Deluxe version, I know the Deluxe board has

1- Firewire
2- Raid (SATA, I think)

How about MAM on the non-delux version? That's really my main concern, because I don't much care about raid and already have firewire. And since I'm on the topic, what is the difference between enabling MAM (which I assume is enabling the PAT technology) and Turbo mode? Isn't Turbo mode just aggressive memory timings?

 

[Rand]

Quoted from THG:

THG crew felt as if it had been overlooked, because actually, the results from the THG lab was what brought this situation to light in the first place.

That's news to me. THG's first look at the P4P800 was the only one of 5 sites I saw that didnt see the P4P800 performing as well as the Canterwood, and they certainly didnt say anything about PAT.
I can think of at least 3 sites that mentioned PAT as a potential explanation for the performance, and one that confirmed it at least 3 days before THG ever mentioned it in their news column let alone at article.
I'd like to see them provide some proof that they were the first to discover it as their published articles indicate the opposite.


The only difference I've seen between the retail and Press P4P800 boards is that the retail release has a few less capacitiors, specifically near the 478pin Socket. That's isnt in of itself cause for worry- Press and full production samples are commonly slightly different as the design/QA is finished they remove the components not deemed necessary or overkill for stability, press samples usually have a few components related purely to BIOS/operation testing that are always removed from retail products.
The two mildly disappointing differences that I've seen are that the AGP signal LED is removed on the Retail boards... or at least all those I've seen.
Also, on Retail boards the NB heatsink is rotated to an angle from the NB positioning itself.
That was probably done due to complaints that the initial positioning on Press samples obstructed installation of larger HSF units. The Retail samples with the slightly rotated HS relative to the NB chip itself arent going to dissipate heat as well though.


I found it quite interesting their Press Sample with the 1006 bios was up to 11% faster then the Retail board.... that is very disturbing. Ironically 10-11% is almost exactly the performance difference between the faster 1005 BIOS and the more recent 1006 BIOS.
They put forth the opinion that PAT is first activated in 1006 with Press/Engineering samples, wehereas it's finally activated in Retail samples with the 1007 BIOS.
I find that to be highly unlikely myself as the press samples with the 1005 BIOS outperform that with the 1006 BIOS.

I wish they had done some tests to see if the Retail P4P800 boards maintained stability at excessive FSB/AGP/PCI bus clocks as well as the Press samples, some investigation into DRAM timings would have been appreciated also.

They reported performance was identical with the latest 1007 BIOS revision however, which is a relief as at least in my experience the 1007 BIOS is the highest performing BIOS... being 1-3% faster then the older 1005 BIOS, and still retaining the expanded VCore range the 1006 bios offered.

 

[LastRide]

I just got my P4P800.There are 2 1007 bios files.One is a beta and the other is final.Which one are they using?.

 

[catss]

Well whatever I really don´t care if the board are not identical, my P4P800 runs like charm with PAT enabled. Theres no overclocking involved like in the MSI case and the performance is stunning, I really don´t understand the commontion about this. If Nvidia and ATI use all tricks to manipulate benchmark thats ok, but if ASUS changes 2 transistors on a board that shocking ? Anyway I see no faulty behaviour on ASUSs site.

 

[ChefJoe]

Wow, my board's still at 1.006 bios. I quess I should be updating as soon as possible (maybe when it's clear that final or beta is what obtained those "near PAT" results. While I understand there's things that get cleaned up, that's a lot of capacitors that got pulled.... capacitors that store energy and even out voltages.... really makes me wonder if there's an overclockability difference as well.

 

[coffeesmurf]

Originally posted by: Slammy1
I read on the ABX forums lots of success with the HyperX 3500. I get a very stable o/c at 260 MHz on a 5:4. Seems in your price range. You might also consider a single DRAM of 512MB with a future upgrade to another. You won't get DCM, but I'm wondering how limiting 256MB will be.

Would it be possible to know what vcore, vdimm and timings you are using for that 260/5:4 ... ?

Thanks !

 

[donfm]

I think what everybody has to look closely at is the difference between performance in the engineering board and the consumer version of the ASUS P4P800. The results are nearly identical in every test. I challenge anyone to actually see a real world difference in everyday computer use between these 2 boards. Yes the engineering sample is missing a few components, but apparently that has NOT impacted performance much at all. I work in Research and Development of a company that makes telephone switching equipment and they are always adding and taking stuff off to increase performance. The fact that a few caps and resistors are missing means nothing really and may actually improve performance at times.

Now in the case of MSI I think that was out and out deception by overclocking the board before the testers even got to it. Even after that they have technical problems galore. Users are afraid to load an upgraded BIOS for fear it will totally disable their computer. I know I was one of them. All manufacturers try to get that extra edge but it is all cheating when you get right down to it.

 

[dwb122]

Man, that was kind of scary news, but the performance issue seems to have been resolved.

The overclocking and stability issue really worries me, though. From that THG article:

"Now, Asus only one more problem left to solve: there are 865PE chips that don't pass the internal Intel PAT qualification tests through speed binning, and there are others that are artificially degraded in order to fulfil market demands. Only the latter type are suitable for stable operation after "reactivating" PAT. Asus wants to ensure this by using strict quality control, and the customers have to put their trust this."

So how much can I trust them? Is there really a good chance that I could end up with an unstable board for OCing?

 

[Davegod75]

asus's site says the p4p800 has raid 0 support only, via the ICHR5 Southbridge. Does anyone know if raid 1 is there or if asus plans to add it via a bios update down the road?

thanks

 

[Slammy1]

Originally posted by: coffeesmurf

Originally posted by: Slammy1
I read on the ABX forums lots of success with the HyperX 3500. I get a very stable o/c at 260 MHz on a 5:4. Seems in your price range. You might also consider a single DRAM of 512MB with a future upgrade to another. You won't get DCM, but I'm wondering how limiting 256MB will be.

Would it be possible to know what vcore, vdimm and timings you are using for that 260/5:4 ... ?

Thanks !

I run everything auto, except my AGP/PCI I run at 66/33. I run my memory by Spd. Turned off memory enhancement (set to auto), but kept Turbo on. I keep my probe alarm set to 60C, and it's never gone off. Prime95 runs 55C after 9 hrs, MemoryGold ran 10 hours no problems. I played some with the memory timings, but I left the voltages alone. I crashed at 270FSB, but 260 is really really great. Even more stable than 200FSB with memory enhancement. I noticed a drop in HD performance at clocks higher than 380/345 on my 9700Pro, but at that o/c no performance drops. Here are some benchmarks, which I may have posted previously.

EDIT: After some light usage, my cpu temp soared to 33C from 29 on start up.EDIT2: Mis-spelled sword, er soared. BIOS=1.005 is supposed to be superior for o/c, 1.006 if you run < PC3200.

Overclock: Manual (260 MHz = 3.14 GHz)
Performance Mode: Turbo
Memory Freq: 320MHz
Memory Accel: Auto
DRAM Idle Timer & Refresh Rate: Auto
Mem By SPD
Graphic Aperture: 128MB
Rage 3D o/c 380/345
PCMark2002
CPU 7613
Mem 9037
HD 1172 (1270 before I installed a bunch of stuff)
Prime95 55-56C for 8 hr
GoldMemory for 10 hours no errors
3dMark01 17967
3dMark03 5965

Sis Sandra
CPU Arithmetic 15309
CPU Multimedia 37174
Memory Bandwidth 10494
File System 32315

 

[xxsk8er101xx]

beta is bad - final is good .

alpha -> beta -> final

alpha means every single line has debugging code.

beta means they fixed some bugs and only neccisary debug code is added

final means debugging code is disabled and has been officaly tested and proven to be stable.

You PC extreme nuts lol don't seem to get it. People don't give a rats butt about performance. They just care about a working computer and thats it. If you wantto overclock and you system becomes unstable - blame yourself.

 

[xxsk8er101xx]

" asus's site says the p4p800 has raid 0 support only, via the ICHR5 Southbridge. Does anyone know if raid 1 is there or if asus plans to add it via a bios update down the road?"

r u even going to use RAID 1?

 

[coffeesmurf]

Thanks for the info Slammy1, i'll give that a try...

 

[Davegod75]

Originally posted by: xxsk8er101xx
" asus's site says the p4p800 has raid 0 support only, via the ICHR5 Southbridge. Does anyone know if raid 1 is there or if asus plans to add it via a bios update down the road?"

r u even going to use RAID 1?

I would not be asking if i did not want to use it

 

[ronnn]

I have the p4p800 and have 2 sticks of kingston khx3500 256mb and a p 2.4c proccessor. My question is .. can I install 2 more sticks of ram with out losing to much memory speed and still oc? Sorry if this has been already answered.

 

[ChefJoe]

Wow, US version doesn't have the 1.007 bios yet.... but the .tw version does

 

[dwb122]

Originally posted by: ChefJoe
Wow, US version doesn't have the 1.007 bios yet.... but the .tw version does

So then what is this?