Dothan @ CT-479 - howto and fixes.

[McNum]

My PM 730 is currently @ 2.7 GHz - 12x225 1.5v.
P4P800 SE
OCZ Performance 2x512 PC3200 @ 225 2-3-3-6
X800XT (modded vivo) @ default settings

3DMark2001 31213
Superpi 27 seconds

The stock cooler is currently holding me back. I don't wanna roast what seems to be a nice CPU

CPU is priming @ 52 celcius atm. Is that too much with 1.5v ?

[macci]

CPU is priming @ 52 celcius atm. Is that too much with 1.5v ?

shouldn't be a problem.
I've noticed that the ASUS P4C800 mobos read quite a bit higher temps from the Dothans than they read from lets say Northwood P4s - even thou the NW is producing a lot more heat and is actually running hotter.
So it would seem that the temp readings w/ Dothan and adapter cannot be directly compared to the readings we got from P4s. The Dothan readings are higher - and more than likely a lot more realistic.

[malfunction!]

yeah, good point.

I also think that the xp90 may be adaptable... will have to get my mitts on one. (the si-97 looks usable too)

C

I was looking at the XP90 / SI97 today (on the net - not in the flesh)... Not sure how the SI97 could be mounted (what with it being a socket A heatsink and all) but the heatsink itself definitely doesn't look like it would get in the way. The XP90 looks like it could be a go-er too... Can't think of anything else other than those and the Zalmans that look like they could work...

[Ychocky]

This seemed like too much fun to pass up.

I'll be doing it in steps. (Purchases atleast)

Ordered the adapter.

Found the dothan ( http://www.ncix.com/products/index.p...ufacture=Intel )

Not sure what ram I have lying about. Probably a gig. EDIT: Yeah, Corsair value select....

Just need the board. Any reccomendations?

I'm certainly looking hard at the P4P800 SE which is about $120 or so here in Ottawa.

( http://www.ncix.com/products/index.p...nufacture=ASUS )

The P4C800-E Deluxe is double the price ( http://www.ncix.com/products/index.p...nufacture=ASUS ) but has a "better" chipset right?

Thanks to all the pioneer's that blew Dothan's before us

Oh and I'm happy to finally join this site, been reading for months now but never had anything to say.

[McNum]

Dothan is definately the way to go. I have a Venice 3000+ too @ 2600 - I like the dottie more

[uclajd]

But Charlie, I don't really see how much closer to ambient temps higher-end air cooling would get you with these chips. I mean they are 27 watts, maybe 33 overclocked. How would a larger heatsink help? At some point, unless you can go sub ambient, I don't see any real benefit, which is why I think non-chilled water would be useless on these.

BTW, I will be hooking mine up today, just trying to get motivated.

just playing with my new ct-479 and p4c800....

and now I can see that you can't just use a normal p4 hs/f with modded clips or longer screws because of the ct-479's fsb jumpers and 4-pin connector on the pcb.
If you have access to a machine shop, you could mod a certain shape p4 hs/f to clear these, kinda' like the asus stock hs/f is built.

Ychocky, welcome to XS forums!

[charlie]

But Charlie, I don't really see how much closer to ambient temps higher-end air cooling would get you with these chips. I mean they are 27 watts, maybe 33 overclocked. How would a larger heatsink help? At some point, unless you can go sub ambient, I don't see any real benefit, which is why I think non-chilled water would be useless on these.

BTW, I will be hooking mine up today, just trying to get motivated.


Ychocky, welcome to XS forums!

yeah, you're probably right, I'm just being a hardware addict!!

C

[runmc]

I'm trying to do the vcore mod on my P4C800 Deluxe.
It's obvious where the one wire is soldered, but I can't see the end of the second wire. I'm almost sure it goes on the end of the resistor that I have marked. Is this correct?


Also what is the best choice bios for the P4C800 Deluxe? I'm using the 1021 atm.

thanks

[uclajd]

K, first post, no problem with P4P800-DX (non-E).

Well, there is one problem - can't adjust voltage with this bios. WTF?



Again, I am using the bios for the regular P4P800, so it reports as such:



EDIT: K, I guess I need to try the P4P800-E bios to get the vcore adjustment?

[charlie]

K, first post, no problem with P4P800-DX (non-E).

Well, there is one problem - can't adjust voltage with this bios. WTF?



Again, I am using the bios for the regular P4P800, so it reports as such:

hmmm... wire mod. VID4 to VSS.... good for about 1.6v or so...

[uclajd]

Looks like from the first post that I can try the P4P800-E bios for vcore adjustment. I'll try that and if I need more juice, I'll go to the wire...

hmmm... wire mod. VID4 to VSS.... good for about 1.6v or so...

[charlie]

when you do, I'd be curious to see how much V you get from the P4C BIOS...

C

[Kamerat]

EIST works fine with the P4GD1. Didn't work on my Celeron M 350 but worked on my Pentium M 730.

[uclajd]

K, I flashed my P4P800-DX (non-E) with the P4P800-E's 1007.003 bios - man, that was scary, since I didn't have a backup of a working bios on floppy or CD, but it worked. Now I have vcore adjustment in bios up to 1.6v.

Highest I can hit on stock air using Clockgen is about 2450, which is kind of disappointing for a 2.0 chip. This board has no mods.

One thing I cannot figure out is how to get the 3:4 divider option.

Also, this thing crashes when I try to change the divider in Windows with CPUMSR and then change the FSB using Clockgen. That is frustrating as hell. Maybe I am using CPUMSR wrong - Do I need to install that driver thingy first?

[Ychocky]

Here are some pics of the hole punched shim.

http://img274.imageshack.us/img_view...05190017tq.png

Nice work, look's like a prospective approach to the shim "problem".

Any board reccomendations?

I'm picking up a cheap Celeron tommorow to flash with (Yay! 2 computers for the price of three ) Maybe I'll buy a cheapo 478 board and give the combo to my mum, I'm sure she needs a gig of ram.

Thank you for the welcome uclajd

[leejsmith]

Most p4p800 boards have vcore droop problems so when running at 1.364v it will droop down to 1.28v ish under load i saw this on my p4p800 after flashing to p4p800-e i could up the vcore to 1.6v but under load it ran at 1.58v.

to get 3:4 memory you need to set the fsb in bios to 133 possible 166 too but i could not get that high with multi of 16 and set memory to turbo. this will set memory to 178. then boot into windows and use setfsb or clockgen to increase fsb keeping 3:4 memory.Using turbo will set your memory to cl2-2-2-5

I tried vid4 to vss but if using p4p800-e bios it is not needed. I also tried adding vid1 to vss to give a possible 1.62v but the p4p800-e still showed 1.6v.


Hope this helps

K, I flashed my P4P800-DX (non-E) with the P4P800-E's 1007.003 bios - man, that was scary, since I didn't have a backup of a working bios on floppy or CD, but it worked. Now I have vcore adjustment in bios up to 1.6v.

Highest I can hit on stock air using Clockgen is about 2450, which is kind of disappointing for a 2.0 chip. This board has no mods.

One thing I cannot figure out is how to get the 3:4 divider option.

Also, this thing crashes when I try to change the divider in Windows with CPUMSR and then change the FSB using Clockgen. That is frustrating as hell. Maybe I am using CPUMSR wrong - Do I need to install that driver thingy first?

[uclajd]

OK, some updated numbers, still without benefit of a decent multiplier:



2001 crashed on launch at 2.753, so 2.7 is about max, which is fine on stock air on an unmodded board at 1.6v. It seems that I can't jump too far at once in Clockgen - if I work my way up, it seems to be more stable?

This is with a stock 6800 GT. Clearly, my numbers are being killed by the low FSB. I really need to figger this multi thing out.

to get 3:4 memory you need to set the fsb in bios to 133 possible 166 too but i could not get that high with multi of 16 and set memory to turbo. this will set memory to 178. then boot into windows and use setfsb or clockgen to increase fsb keeping 3:4 memory.Using turbo will set your memory to cl2-2-2-5

Thanks for the tips Lee, I will try this.

I want to get this thing fully squared away before I give it the phase change treatment.

[Malachia]

nice uclajd, you have done boot block for flashing these bioses???

[uclajd]

Here are some pics of the hole punched shim.

http://img274.imageshack.us/img_view...05190017tq.png

Nice approach.

I am going to try something similar with some mounting pads or tape I just picked up from Home Depot when I do the phase thingy.

[leejsmith]

An XP 90 will work after you mod the ct-479 to remove the power socker and solder the wires/jumpers direct.

I managed to get the xp 90 to fit with the standard mounting clips but it was a lot of pressure and i chipped the corners of the dothan core. I made some adapters out of u shaped aluminium one end went into the p4 bracket the other had a hole for a bolt to go through and i used the bolts to secure the xp-90.



this also worked for my cascade water block but it was no better than the xp-90.



I think the zalman will just need some longer bolts to fix the cooler to the brackets in the sockets should be easier than the xp-90.

with the dothan at 133x16 2128 1.48vcore it idles at 25c and load shows 33c while system shows 25c. Not sure how true those temps are as the p4p800 boards are known to show lower temps than other boards. This is with a 90mm zalman fan at 7volts.

I was looking at the XP90 / SI97 today (on the net - not in the flesh)... Not sure how the SI97 could be mounted (what with it being a socket A heatsink and all) but the heatsink itself definitely doesn't look like it would get in the way. The XP90 looks like it could be a go-er too... Can't think of anything else other than those and the Zalmans that look like they could work...

[Malachia]

uclajd maybe your problem is the fsb400 cpu....everybody with that cpu cannot pass even 200...

[uclajd]

nice uclajd, you have done boot block for flashing these bioses???

Not for the stock P4P800-DX-to-P4P800-1021.003 (at this point, I was running a 3.2E on the board), I just used the current Windows flash utility for that. But for 1021.003-to-P4P800-E_1007.003 flash (now with the 760 onboard, and wanting the vcore adjustment), of course the utility gave me the, "hey dumb bunny, you can't do that message," so I had to go the DOS route with the older afudos utility that will let you flash different bioses. For that, I didn't take any chances and did the "afudos /i1007.003 /n /pbnc" route. And yes, that scared the shiat outta me.

[uclajd]

this also worked for my cascade water block but it was no better than the xp-90.

Naturally, LOL, cause you are already near ambient with the xp-90, lol. I think water is useless on these things, unless it is chilled.

Nice work on the brackets though!

[leejsmith]

cpumsr will change the multi for you once your in windows but as your not setting fsb to 200 in the bios you might have the same problem i have gettings the fsb over 168 was impossible no matter what multi i used.

edit: just noticed from your post yes you have to install the driver for cpumsr to work.


My best 2k1 score was 22000 with a 6600gt running at 560/1150 and the dothan at 2590 162x16

I hate my main p4 system after playing with the dothan. I spent loads water cooling it keeping the noise levels as low as possible pushing it as far as i can and then i play with the dothan and it beats my 3d mark scores super pi scores and is almost silent.

OK, some updated numbers, still without benefit of a decent multiplier:



2001 crashed on launch at 2.753, so 2.7 is about max, which is fine on stock air on an unmodded board at 1.6v. It seems that I can't jump too far at once in Clockgen - if I work my way up, it seems to be more stable?

This is with a stock 6800 GT. Clearly, my numbers are being killed by the low FSB. I really need to figger this multi thing out.

Thanks for the tips Lee, I will try this.

I want to get this thing fully squared away before I give it the phase change treatment.

[Malachia]

then first to buy a p4c i will try the p4p dlx with -e bios...nice...