2.4C, P4P800 non-deluxe @ 266fsb Final

[Thor86]

P4 2.4C L311Axxx
Corsair TwinX3200LL (2 x 256)
Asus P4P800 non Deluxe rev1.02 (1007 bios)
Thermalright SLK900U w/ Panaflo 92mm H1A

1006 bios
250fsb, 1.525v, MAM=Auto, 3:2, 2.5-3-3-7-8, sisoft 4565/4526, aida32 3876(read)/1189(write)
266fsb, 1.600v, MAM=Auto, 3:2, 2.5-3-3.7-8, vdimm=2.75, vagp=1.5, prime95 two instances passed 18+ hours (f/l 1.57vcore, idle 1.66vcore)
267fsb, 1.600v, MAM-Enabled, 5:4, 2.5-3-3-7-4, vdimm=2.75, vagp=1.5, prime95 two instances failed 1 instance 10+hours (f/l 1.57vcore, idle 1.66vcore)
269fsb, 1.650v, MAM=Auto, 3:2, 2.5-3-3-7-8, vdimm=2.75, vagp=1.5, prime96 two instaces failed (f/l 1.57vcore, idle 1.68vcore)
269fsb, 1.600v, MAM=Auto, 3:2, 2.5-4-4-8-8, vdimm=2.75, vagp=1.7, prime96 two instaces failed 1 instance 32mins
275fsb, 1.525v, MAM=Auto, 3:2, 2.5-3-3-7-8, sisoft 5012/5090, aida32 4243(read)/1263(write), prime95 failed
285fsb, 1.600v, MAM=Auto, 3:2, 2.5-4-4-7-8, vdimm=2.75, vapg=1.8, max bootable into windows
286fsb, 1.650v, MAM=Auto, 3:2, 2.5-4-4-8-8, vdimm=2.85, vapg=1.8, bsod

1007 bios (Performance Mode)
200fsb, 1.525v, Performance Mode=Turbo, MAM=Enabled, 1:1, 2-3-3-6-8, vdimm=2.75, vagp=1.5, prime95 two instances passed 1+ hour (f/l 1.54vcore, idle 1.58vcore)
sisoft 4917/4928, aida32 5017(read)1578(write) - memory timings default to SPD (2-2-2-5) according to cpu-z when using Performance Mode set to Turbo.
200fsb, 1.525v, Performance Mode=Standard, MAM=Enabled, 1:1, 2-3-3-7-8, vdimm=2.75, vagp=1.5, prime95 two instances passed 1+ hour (f/l 1.54vcore, idle 1.58vcore)
sisoft 4735/4753, aida32 4690(read)1441(write)

(5:4 memory scores)
266fsb, 1.600v, Performance Mode=Standard, MAM=Enabled, 5:4, 2.5-3-3-7-4, vdimm=2.75, vagp=1.5, prime95 two instances passed 4+ hours (f/l 1.58vcore, idle 1.66vcore)
sisoft 5625/5543, aida32 5066(read)1471(write)

(3:2 memory scores)
266fsb, 1.600v, Performance Mode=Standard, MAM=Enabled, 3:2, 2-2-2-6-4, vdimm=2.75, vagp=1.5, prime95 two instances TESTING (f/l 1.58vcore, idle 1.66vcore)
sisoft 5154/5152, aida32 4730(read)1433(write)

266fsb, 1.600v, Performance Mode=Standard, MAM=Enabled, 5:4, 2-3(ras to cas delay)-2-6-4, vdimm=2.75, vagp=1.5, prime95 two instances TESTING (12+ hours so far!) (f/l 1.58vcore, idle 1.66vcore) sisoft xxxx/xxxx, aida32 xxxx(read)xxxx(write)

266fsb, 1.600v, Performance Mode=Standard, MAM=Enabled, 5:4, 2-2-2-6-4, vdimm=2.75, vagp=1.5, prime95 two instances 1 failed (f/l 1.58vcore, idle 1.66vcore)
sisoft 5707/5695, aida32 5238(read)1578(write)
267fsb, 1.625v, Performance Mode=Auto, MAM=Auto, 3:2, 2.5-3-3-7-8, vdimm=2.75, vagp=1.5, prime95 two instances failed (f/l 1.55vcore, idle 1.63vcore)
267fsb, 1.625v, Performance Mode=Turbo, MAM=Enabled, 5:4, 2.5-3-3-7-4, vdimm=2.85, vagp=1.7, prime95 two instances failed (f/l 1.55vcore, idle 1.63vcore)
270fsb, 1.625v, Performance Mode=Standard, MAM=Auto, 3:2, 2.5-4-4-8-8, vdimm=2.75, vagp=1.7, prime95 two instances failed (f/l 1.57vcore, idle 1.63vcore)
270fsb, 1.675v, Performance Mode=Standard, MAM=Auto, 3:2, 2.5-4-4-8-8, vdimm=2.75, vagp=1.7, prime95 two instances tested 15mins (f/l 1.6vcore, idle 1.71vcore)
275fsb, 1.700v, Performance Mode=Standard, MAM=Auto, 3:2, 2.5-4-4-8-8, vdimm=2.75, vagp=1.5, HyperThreading disabled, prime95 failed (f/l 1.63vcore, idle 1.68vcore)
275fsb, 1.675v, Performance Mode=Standard, MAM=Auto, 3:2, 2.5-4-4-8-8, vdimm=2.75, vagp=1.5, HyperThreading disabled, prime95 failed (f/l 1.58vcore, idle 1.65vcore)
287fsb, 1.725v, Performance Mode=Standard, MAM=Auto, 3:2, 2.5-4-4-7-8, vdimm=2.75, vagp=1.7, HyperThreading disabled, prime95 failed (f/l ?.??vcore, idle 1.71vcore)
290fsb, 1.700v, Performance Mode=Standard, MAM=Auto, 3:2, 2.5-4-4-7-8, vdimm=2.85, vagp=1.8, HyperThreading disabled, prime95 failed (f/l ?.??vcore, idle 1.68vcore)
290fsb BSOD

 

[Quackmaster]

Nice stuff! I got a p4p800 deluxe (product code-SL6WF, fpo/batch-L309A586) and I've come up with similar results so for ( I, too, still haven't finished tweaking it out):

P4 2.4c
Kingston Hyper X 3500
Asus p4p800 deluxe rev1.02 (1005 BIOS stock, still haven't upgraded...yet)
stock heatsink. My tower however has my Enermax PSU right above and next to the cpu socket. Additionally my case has a 12cm fan located at the rear of the case within inches of the CPU. My temps do change drastically if this 120mm is turned off. Temps are high tho. Today is the first cool day we've had in awhile here in LA so I'm benching/clocking it in between working. Ambient is 29c here (boy it's nice to be back in the 80s! Nice and cool ! ), but my temps are 36 idle, 52 load !! That's at the highest overclock I've taken it to thus far ( 285fsb), but I'm gonna see what else I can squeeze outta this baby after (IF) it finishes it's current Prime/Sandra/bench run. Normally I'll keep it much lower (3.0-3.2GHz) because of ambient temps being in the 30c and higher range.

Also, my temps are coming from a cheapo thermometer from Home Depot, Asus Probe and Motherboard Monitor. So they are likely quite a bit innacurate , but a ran the cpu for 9 hours last night/today doing prime and sandra and 3dMark loops (simultaneously) and it never crashed once at 280fsb 2-3-3-8-8 cpu @ 1.6volts RAM at 2.65v. AGP stock voltage. Currently it's running through the benches at 285 and is looking good. Temps have only changed 2 degrees celsius from 3.18GHz to 3.42GHz. I'm going for 290-300 now!

 

[Thor86]

Updated.

 

[Duvie]

Hey Prime96 huh???


Looks good!!! Got a 800mhz OC out of it and running DCDDR pc3500 level...Beats my 3.2ghz!!!! PLus HT...Do you have that enabled???

 

[Quackmaster]

Originally posted by: Duvie
Hey Prime96 huh???


Looks good!!! Got a 800mhz OC out of it and running DCDDR pc3500 level...Beats my 3.2ghz!!!! PLus HT...Do you have that enabled???

Oh yes.. I love this hyperthreaded stuffs.

Thor, I've heard that the 1006 and 1007 BIOSes fix the cpu hitting a wall at 270fsb (cpus can OC above ~270, but get stuck if on that MHz exactly. Can you confirm this on yours?

I'm gonna bench tonight while I sleep, but 290fsb was a no go...:sad; I think it's the RAM (long story why) Still, believe I've got a monster cpu on my hands, 'cuz it's running now (I've backed it down) at 283fsb (3.4GHz !!) 2-2-2-7-8 2/3 ratio 1.56volts (with Hyperthreading of course ), RAM at 2.65volts and AGP default. My PCI/AGP buses are at default as well (33/66).

I bought this RAM cuz it was cheap ($69/256MB) and was only US11 over the cost of PC3200. I also expect later this summer or fall for us to get the RAM that will do low latencies at ~250MHz+ (270? drooool ) So that's also why I went cheap now. It won't hit it's rated speed tho!! (216 MHz) grrrr :mad; Still, a killer cpu vs weak-a** RAM? I'll take that trade.... for now.

 

[Thor86]

Duvie
Hehe, I was pretty tired when posting as you could tell. Yes, HT is enabled, and two instances of Prime95 are still running 8+ hours at 267fsb. If you only run one instance of Prime95, only 50% of the cpu is utilised. HT is the bomb, no delays on the desktop when loading apps even under full cpu load. I suppose the 250+fsb helps with that too! Will test 268fsb which I think is the max for this cpu, until I flash to 1007 bios and see if I can reach higher.

Quackmaster
Congrats on your cpu, that's one killer processor! Anything over 275fsb is IMHO AWESOME! I will try out the 1007 bios tonight as well after work and see if that helps me past 269fsb. One instance of Prime95 always bombs out at 269fsb at the highest voltage (1.65) that I give it using 1006 bios and even at 3:2 and most relaxed settings as you can see in my first post. Yeah, can't wait until DDR600 comes out.

 

[Thor86]

Tried the 1007 bios, and still cannot Prime95 past 267fsb without erroring out. The only differences I found with the 1007 bios from 1006 is that the vcore overvolting seems less and the Performance Mode is back - which is useless in anything but 1:1 memory ratios. Will try 1 stick at a time and see if that may let me Prime95 past 267fsb.

Edit: May even try and disable HyperThreading for max fsb overclock.

 

[cjsketchy]

I've almost got the exact same setup as you - different ram (ocz pc3500) - but I get almost the exact same results as you. Neat

I didn't know about the 270fsb limit - I tried it a few times and had the clear the cmos. So I just assumed I couldn't go any higher with my cpu. (With 1005, 1006 and 1007 bios). Guess that's why. I should try something higher than 270.

(/newbie overclocker out)

 

[Thor86]

Originally posted by: cjsketchy
I've almost got the exact same setup as you - different ram (ocz pc3500) - but I get almost the exact same results as you. Neat

I didn't know about the 270fsb limit - I tried it a few times and had the clear the cmos. So I just assumed I couldn't go any higher with my cpu. (With 1005, 1006 and 1007 bios). Guess that's why. I should try something higher than 270.

(/newbie overclocker out)

Could be the motherboard, could be the cpu, unless I have two of each or different motherboard, then who knows?

 

[oldfart]

Looking good so far. With that ram, you should have no trouble going to 5:4 @ 266 FSB.

 

[Thor86]

Thanks OF!

Looks like this is all she wrote!

266fsb, 1.600v, Performance Mode=Standard, MAM=enabled, 5:4, 2.5-3-3-7-4, vdimm=2.75, vagp=1.5, prime95 two instances passed 4+ hours (f/l 1.58vcore, idle 1.66vcore)
sisoft 5625/5543, aida32 5066(read)1471(write)

Now testing my 2.6. Should have a post later tonight.

 

[Tom]

Looks great ! btw, how much did your memory scores go up going from 3:2 to 5:4 ?

Like you I think better ram at lower prices will come later this summer so Im' getting along with old ram for a while.

 

[Quackmaster]

Hey peeps!

Got an update for ya's, interesting stuffs.

Ok, so like most of us I upped the voltage for stability at the expense of heat. Before I ever had done this I had first done what I always do, and that is to drop the voltage first and see what I can get outta my cpu. It did 3.0, but then crapped out after 258fsb. So far, so good using lower voltage.

Then I did the usual OCing and got up to the 3.4 I posted earlier using cpu@ 1.6voltage and RAM at 2.65voltage (BIOS setting). I've been running it since last Friday now at 3.0GHz continuously, and since Sunday at 3.2GHz up to 3.4, ya with me? This included basically running it around the clock doing 3dMark loops, Prime, Sandra, DivX encoding (it's easy to see when there are errors this way. I just watch the movie! ). So basically I've been "training" (or "punishing" :evil: ) my cpu to run hard all the time.

Well this afternoon I backed off to cpu voltage and kept dropping it, testing.... dropping voltage, testing. and to make this short(er) I've got it running at 1.525volts @ 3.4 quite well. :Q It seems stable, but I need more long term testing. If I run at 3.3GHz (275fsb) it's bulletproof.

I'm gonna continue testing, but basically I've run the CPU into the ground at higher voltages, and now am able to run it at (or nearly at... still need more testing) a much lower voltage. This trick has worked for me before on an old AMD K6-2 and on a P3 500e FC-PGA, and also a Tualatin (P3) 1.2GHz.

So beat up these cpus 24/7:clock: and give it a try. Maybe you're got a 3.6GHz barnburner waiting to be released!

 

[ErikaeanLogic]

good work! that's the kind of news I'm hoping for, got 2 more P4 C's on the way

 

[RandySavage]

that sandra score is sweet... I believe using the performance modes on the ic7 have killed my raptors and memory for some reason... @ 270 fsb 5:4 800 nbs, I get 5100/5100 on mem bandwidth on sandra. Before I began my overclock, I got 69k on the hd bench and now I get 62k... doh!

I like the 2.4C... I know nothing about overclocking and I am sitting at well over 3 ghz...

but great OC man! Smokin!

 

[Quackmaster]

Oh, btw- I forgot to mention that you should all take any rotten eggs, bird droppings, are other disgusting projectiles and hurl them at me now :Q

The reason is because I am an idiot! I forgot to mention this in the last post, but bascally I mentioned earlier how my Kingston was crap and I was having to run at the 2/3 ratio because I couldn't get it stable at 216MHz. Well that was me having too many options and too many choices to play with. I am now running like this-

2.4c @ 275fsb =3304MHz. cpu @ 1.5voltage(stock setting in the BIOS) Asus probe reads 1.536volts. temps (from Asus probe) are now at 95F (36c) idle, 124F load. The BIOS reads these temps as slightly higher and that pisses me off because I know the temps drop rapidly before I can get to the monitoring screen in the BIOS. They drop hella quickly in Windows just while watching Asus Probe (henceforth to be called aspro lol ) after I exit Prime95 or Sandra. Obviously I NEED a digital doc5 or somethin' like dat. Oh, and I'm still on the 1005 BIOS. I'll back it up in case anybody needs it since Asus deleted it off their site (dunno what's on my Asus cd, gotta look).

Kingston HyperX @ 4/5 ratio with 2.75volts(BIOS setting). fsb is @ 275. I can't get this to run above 279 w/ the 4/5 ratio, but now can run it at 220MHz instead of the 189MHz I was running with the 2/3 ratio. hmmmm...... So I'm giving up 100MHz for way faster memory bandwidth. The reason this all works now (also the reason I deserve to be pelted with cat poop and chicken feed) is because I got confused between, work, clocking this 'puter, and trying to manage my other responsibilities. As a result I forgot to try my RAM with the BIOS set to spd.. DOH! *smacks head against wall*. This was something I did initially up to the 3.0GHz mark, but totally forgot about after that. geez. I still can't believe I was that dumb. grrrrr. Anyways, thanks to the "RAM by spd" BIOS setting my 3.3GHz Sandra marks look like this (averages of 5 runs) -


5641 interger (buffered of course)
5675 floating point

unbuffered marks are more useful IMO so I like seeing these more-

3803 int
3798 float


I'll post 3dMark in a minute or two, I would think the extra memory bandwidth ain't gonna help as much as the extra 100-140Mhz from the cpu, but we'll see. I dunno.

Edit- Ok, these are my new scores (average of two runs)

16,873 at default 1024 w/ no AA etc.

So they dropped since I downclocked the cpu ~140MHz. But I can still make that up later with software optimizing (haven't tried any yet. Too busy playing with the hardware).

Even with the decrease in clock speed, I'm still not sure what I like better (decisions, decisions...hmmmm ), higher clocks and higher temps, or lower clocks with better memory scores. Since I'm in LA and all this clocking madness has been going on while we've been having unusually cool weather, I think it's likely that this 3.3GHz is where I'm gonna keep it (until tomorrow anyways ). I'm gonna see how low I can take the cpu voltage before it becomes unstable. Then, I'll probably bump it back up next week (I'm leaving town this Thursday for the weekend) and see what the final max is that I can do on this cpu with stock cooling.

Hey Thor, got any new info? I'm curious to see what you get since you're up there in the nice and cool Canadian landscape. I'm also curious (if you try it) to see how the run-it-hard-overvolted-then-undervolt-it method works for you (anyone else out there try it?). So what's up postmeister? I'm splittin' tomorrow. Don't make me wait all weekend!! whaaaaw

 

[Thor86]

Dead Parrot Sketch
I'm not sure about 3:2 memory scores, but when I put the 2.4 back into my system, I will bench them at that ratio and update this thread for ya! I would assume they will be lower than at 5:4.

Quackmaster
That's awesome stuff you got going there! I wish I had gotten me a deluxe board, but I wouldn't think they were much different other than the price/features which I didn't need. May have to try and go back using 1005 bios and retest my 2.4. You may be onto something there about some period of "burn-in" with these new processors. I think running them for 24+ hours full load at stable conditions may in some way contribute to higher/stable overclocks afterwards. BTW, the weather up here is getting warmer, so my ambient temps are about 5-10 degrees higher than usual. I'm not worried though, I've got A/C, but haven't bothered turning it on yet - its not hot enough.

ErikaeanLogic
Looking forward to your OC/Testing!

RandySavage
Thanks, I really like the 2.4Cs, the higher the FSB the snappier the desktop seems. No delays to load anything up, or switching between tasks, even under full load. I can even "feel" the difference between my 2.4@266fsb and my 2.6@256fsb.

Well, off to post my 2.6 info.

 

[cjsketchy]

I tried going past 270 and it still won't boot...

Going to get another 2.4c today, and I've got an IC7 coming soon, so we'll see which part is the weakest link

 

[Thor86]

Originally posted by: cjsketchy
I tried going past 270 and it still won't boot...

Going to get another 2.4c today, and I've got an IC7 coming soon, so we'll see which part is the weakest link

Have you tried using the 1007 bios? I couldn't boot past 284fsb with 1006, and then the 1007 let me boot up to 290fsb. I blue-screened when trying to boot to 295fsb.

Let me know how the IC7 goes for you. I may just look into one of these as well.

 

[anomaly]

What made you choose the P4P800 over the IC7, for those who made that decision?

 

[Thor86]

Originally posted by: anomaly
What made you choose the P4P800 over the IC7, for those who made that decision?

Price/availability, but price mostly.

 

[cjsketchy]

Yeah, I tried the 1007 bios, with the same result. I played around with just about every setting in the bios with no luck. I think with these boards we're going to see a fairly wide range of OC results because of the optimizations that asus is doing with the 865PE chipset and speed binning. From what I gather, these are the canterwood parts that didn't make the cut - so some of them are going to be on the high end of that performance spectrum, and others will be stinkers that won't do much more than stock speeds. But with a canterwood board you know you're getting a chipset that met some higher standards. So the OCing results should be consistently higher with a good CPU. With the P4P800 it's the luck of the draw. That's what I think at least.

Basically the reason I got the P4P800 is because I got the itch to try out a C chip, the board had just arrived at a store close by, and it was reasonably priced. I couldn't find an IC7 anywhere locally and I wanted it up and running asap If I could find an IC7 in town then I probably would've gone for that instead.

 

[Thor86]

Originally posted by: cjsketchy
Yeah, I tried the 1007 bios, with the same result. I played around with just about every setting in the bios with no luck. I think with these boards we're going to see a fairly wide range of OC results because of the optimizations that asus is doing with the 865PE chipset and speed binning. From what I gather, these are the canterwood parts that didn't make the cut - so some of them are going to be on the high end of that performance spectrum, and others will be stinkers that won't do much more than stock speeds. But with a canterwood board you know you're getting a chipset that met some higher standards. So the OCing results should be consistently higher with a good CPU. With the P4P800 it's the luck of the draw. That's what I think at least.

Basically the reason I got the P4P800 is because I got the itch to try out a C chip, the board had just arrived at a store close by, and it was reasonably priced. I couldn't find an IC7 anywhere locally and I wanted it up and running asap If I could find an IC7 in town then I probably would've gone for that instead.

I think you are right on the money with this conclusion. Some P4P800s will do better than others due to the chipset binning by Intel/Asus.

 

[Quackmaster]

I agree with you two, I think the Canterwoods will pan out to become the better boards in the long run. If not now, then definitely when we hit the 2nd generation of these boards sometime later this summer/early fall.

I do wonder about one thing tho...

Intel has had all their fabs running on the .13 process for a long time now, and already has two of them transitioned(ing) over to .09. Given their high level of quality control (high chip yields), Intel's been cranking out the .13 process (which has been a long time now, since the just after the Pentium 3 days and the P3 1.13GHz debacle), long enough with high enough yields, that I'm guessing (more on that in a sec) they actually have a surplus of "Canterwood approved" chips. Not to mention all the P4s that have been great overclockers percentage-wise. Think about it, the 1.6GHz had an average overclock of 2443MHz (I'm getting this data from the overclockers database), the 2.0 had an average of 2773MHz, now we've got all the new 800fsb "c" chips getting massive overclocks.

The standard used to be that if you could OC a chip 30%, that was great, but if it could go to 33% or even higher, then it was golden and deserved hall of fame praise. But ever since Intel got this .13 process under control (about the time we saw Northwoods start to appear), they have just been cranking out chips that OC like mad. So why would their manufacturing be any worse on chipsets.....? It wouldn't!!

The reason I know this is I've got a good freind who works for UMC (the Tawainese company that make ATIs chips) which is a sister company to TSMC (who was slow getting their .13 straightened out, which led to Nvidia's 6 month delay of NV30 aka Geforce FX). To make this long-a** story short. I've held a 12" wafer in my hand (so pretty ) while my buddy explained to me how fabs tighten up QA as they refine their process. Given that premise, and the fact the Intel OWNS all their fabs, I'm betting that we will continue to see high-performing Springdales with only a few lemons.

Intel just cannot hoard the good chips and make more of the bad chips, it doesn't work like that, and besides, that would cost a TON of money (guess how much a 12" wafer costs raw?) and leave Intel with a ton of inventory. So that puts the Canterwood vs Springdale battle in the mobo makers hands. Do they want to sell more of the Canterwoods with their higher profits? Or do they wanna just sell boards, period? The computer market is already being projected to be dropping off in sales again this quarter and next. Intel, the tawainese mobo makers, and even the RAM companies are all going to see lower sales (that's why all the DRAM makers are hoarding their chips right now. So they can jack up the price later, but that's another topic). So what better way to counter that than with (relatively) cheap boards that kick buttocks? For every one of us geeks in the hardware community buying a $$$$ IC7-G, there will be 4, 5, or maybe 10 who will buy a Springdale. Remember, Canterwood is a workstation board costing more than most people wanna pay. Joe Blow would rather buy a faster processor than pay for some fancy-shmancy mobo with features he doesn't need/want.


So whaddya think fellas, does that sound feasible/reasonable? Or have I had one too many double-shots?

I bet in the long run, Canterwoods (on average of course, there will always be exceptions) will have a slight edge of no more than the 3-5% spec they were originally claimed to have. And that's at best. Until, Intel makes another die shrink on their chipset process, I think we're gonna see a lot of quality chipsets from their budget lines, whether it's Springdales, or i845s, or some Grantsdale knock-off.

 

[Quackmaster]

Oh, and to answer your question anomaly, I originally planned on buying a Canterwood way back in March/Feb somewhere around there. I wanted the extra 5%. But once I saw the features that were being offered for the price, I decided to play "wait and see which mobo turns out to be the real "sleeper" or "killer". I've been through this before with the BH6, BE6, BE6-2 and all that AMD KX133, KT133 etc etc BS. My way of thinking is if two boards have a similar feature set and are both stable, then I'll often buy the cheaper one and put the saved money towards a better cpu.

I also am an Asus whore. But I have bought Gigabyte, MSI, Soyo, Abit and FIC to name a few. So when I initially decided to get a springdale because the Canterwoods weren't performing THAT much better (if at all), it was a contest between the Asus, Abit is7-e, and the Albatron (looks, price, and features. Plus Anands review). But I was wary because of the whole MSI situation, and Abit had a few years there were their quality control sucked, so I waited patiently, but pretty much was set on the Abit, then the p4p800 started blowing up and I knew I had found the "sleeper".

Bad thing is I'm building a hand built case that is quite the looker, and this Asus board is pretty in it's utilitarianism, but it ain't half as cool looking as the Albatron, or even that hideous DFI glow-in-the-dark-LAN-Party-crap. Oh well. Guess I'll have to wait for the 2nd round of new designs for 865/875.