PC-DL 800 FSB Mod

[mateo]

Wow, you are a real man. Only real men break pins off of a xeon to get a better clock... Amazing. I invite you to join FAPLOKAP: Federal Association of Pirates, Luberjacks, and Other Kick-*** People.

Say what? We're not the [H] or anything, not like we're really looking for that. Our purpose for participation here is to experiment and push the envelope of performance potential, and this is merely another means of doing that. Modding is modding; if a concept works, which, in a simple case of 0-1 logic like this there's no reason for it not to, why not try it? This isn't like trying to mod an ATI card where there's only a chance for it to work- this is clear-cut modification straight off of Intel's own docs on hardware known in a uniprocessor system to be capable of handling such a boost. And, too, you can insulate pins w/out snapping.

 

[Dc5e]

Polaris, tell us how it comes out, i'm planning to get a dual xeon system myself next weekend

 

[Polariz^]

Will do matey, i'm having a slight postage problem (I missed the courier - bugger), so will be picking up the motherboard tommorow. Heatsinks and PSU on their way to me now, hopefully will be fully built by thursday and will let you know of it's overclocking potential.

I did try to locate m0 CPU's however could only find one of them. I'm in the UK, however even though I picked up my CPU's from Ebay, www.matrixthree.co.uk is a British site that will ship to anywhere. Even though I didnt buy from this place, I did find the communication very good. Obviously the benefit is however that you get to select which stepping CPU's you can get. Reading from 2CPU.com however, only 2.4's or lower work with this mod - O/Cing starts at the buying stage remember!

Sorry for the advert if it's not allowed; i'm in no way connected to this site, but if it helps OCers then go for it .

 

[Fieldski]

Reading from 2CPU.com however, only 2.4's or lower work with this mod - O/Cing starts at the buying stage remember!

That's not true, I do believe that hitsquad over at the pc-dl user site did accomplish the "moosemod" with his 2.66's...Although we are not clear which mod he did....The "Moosemod" or the BSEL


http://www.datamine.tk/forum/forum_posts.asp?TID=17&PN=1

fieldski

 

[Polariz^]

Cheers mate.

 

[Polariz^]

Ok I dont think the moosemod is working for me - crap...

I'm using 2x D1 stepping Xeons (Identical), but when I apply the moosemod only one of the CPU's shows up in BIOS (Hyperthreading enabled). And also shows only two CPU's in Windows.

Any ideas? I'm not best pleased : /

 

[Polariz^]

Ok ive put the jumpers back to what they were/ Is there any reason for this to happen? Ive done no voltage mods or anything, but the chip has reappeared now im back at 533 so luckily the chip isnt dead.

Would it be something to do with the other bus speeds? I havent found any locking options yet. Are there any?

Thanks.

 

[Polariz^]

Right I just noticed i'm running 1003 BIOS, is this a contributing factor?

I bumped the FSB up to 165 and it booted up ok, but the soundcard had stopped working, probably due to the bus speeds.

Bugger

 

[Fieldski]

What revision are you running...1.04 or 1.05

 

[Polariz^]

1.05, unlike most other people...

Any suggestions? A BIOS upgrade? Or perhaps one of the chips is being undervolted? I tested them first using 12x200 to reach the same speeds however I still ended up with the same problem - one of the CPU's is just missing : /

Thanks matey.

 

[Polariz^]

*Bump*

Somebody help me please!

One other thing us that I have a backup plan if the moosemod doesn't work, is that i'd like to do this by fsb. I used clockgen to increase the FSB, however the AGP and PCI buses were also increased by a great deal - how do I go about locking these buses? (If possible).

Would really appreciate the help

 

[disk11]

Try a newer BIOS. Either one should work fine.

 

[Audioaficionado]

Try a newer BIOS. Either one should work fine.

Don't use that Asus windows updater unless you want to end up ordering a spare BIOS chip from Asus. Stick with the floppy method even if you have to temporarily hook one up.

 

[Polariz^]

Ok i'veupdated to BIOS 1005, so now when I perform the moosemod I simply get no POST.

Could this be down to crappy chips? Or not enough voltage? I'm not all that enthusiastic about performing the Wire mod...

 

[Polariz^]

I've used clockgen to o/c the chips now that i've figured out how to lock the AGP and PCI buses, but it is unstable at a rather modest overclock to 3GHz/170FSB/217 Memory.

Voltage im guessing - CPU-Z reports the Voltage for both CPU's at 1.461V-1.471V...

Does it matter then, how thick the piece of wire is for the volt mod? As long as it is not to thick as to stop the pins from entering the socket? Or must I use something specific?

 

[DaveB]

I used single strands of wire from some generic speaker wire I had laying around. they worked perfectly and even survived several CPU removals. Do the single-wire 1.6V mod to get the actual core to 1.56V or so.

The MooseMod really only seems to work with 2.4/533 Xeons. Both M0s and D1s work fine, I even seem to remember one guy succeeding with C1s. But I really think 2.66/533 and anything higher are not gonna work with the MooseMod. I originally had two 2.66/533 C1 (SL6GF) that did 3.3 GHz at the default core and went to 3.4 GHz with the Vcore mod. I sold them prior to Mooseifert's discovery so I couldn't try them with his mod. But ClockGen can be setup to OC to whatever you want when Windows loads anyway. Without the MooseMod, however, you pretty much need some BH-5 or BH-6 DDR and the Pomona SMD Grabber solderless VDIMM mod to get over 200 MHz. The mod is easy, virtually risk free, and you don't have to solder anything on the motherboard. Remember that to get the AGP/PCI lock, the FSB must be set at 134 MHz or higher in the BIOS. If you leave it at 133 MHz, the AGP and PCI clocks will increase right along with the FSB.

 

[Polariz^]

I am an extremely happy man! I did the 1.6V socket mod and it's working great - phew!

Thanks guys, you're Gods.

I'll post up some OCing results and stability tests very shortly.

 

[Polariz^]

Ok the results werent great at all, but i'm happy enough.

The most stable I can get it using 1.57V is at 2.67GHz / 205FSB / DDR400. The bus increase is enough for a bit of an increase in performance, as well as a 250MHz OC. I think the problem all along has been down to one of the chips being a bit rubbish. I did buy them second hand, and obviously they are D1 and not M0 so I was expecting anything fantastic...

Would have been nice to at least hit 3GHz though

 

[NeoGeo]

I think the problem all along has been down to one of the chips being a bit rubbish.

What's wrong with plugging in and testing one CPU at a time? You would get to see how far each goes and if one of them is holding you back, no?

 

[Polariz^]

It took me an hour to get the heatsinks off - it's staying the way it is

I did consider doing this now, but i'll do it when I can afford to buy another one, so for the moment i'll stick with it.