Bringing the Pentium-M to the desktop

[caater]

Which is the newer chip in terms of when it was made?
Scott.

both are made the same week at the same plant
7514 - y2005 week14..

 

[Jimbob7]

Very strange! See what it can do FSB wise. My Dothan doesn't like much over 2.7 even there it's a bit flakey. But it can handle a nice FSB as far as 875 chipset goes.

That second 780, is a bit of a sucka, atleast you have one amazing chip . I guess batch date week etc, doesn't apply to Dothans, or this is just an extreme case lol.

Scott

 

[Sjaak]

W00t! Dothan twins!

Maybe intel forgot to recalibrate the temp sensor in one of em.

 

[TechnoButt]

Chip: 1.6 533FSB
Motherboard: P4P800-E DLX
RAM: 2x512MB OCZ EL 4200 (TCCD, I think)
Videocard: Leadtek AGP 6200@6600+
Cooling: CT479 Stock
Current Speed: 2400Mhz (12x200)
Max FSB / CPU clock: 230Mhz / 2750Mhz
Current Voltage: 1.45v
S-code / stepping / pack date: SL86G / B1 / unknown

I traded my SL7GL for an SL86G (both OEM pulls). This one is a bit stronger. It ran super pi 1M @ 2730mhz (12x228) with the OCZ EL 4200 at 285mhz (DDR570) in 27S. All on stock CT479 cooling.. I've got an XP-90 laying around.. I may try to mod for it.. but this is my fiance's rig.

Main rig: X2 [email protected] on CNPZ7700Cu w/EL 4800 OCZ and BFG 7800GTX.. we've got the fastest web browsing, bill paying, email reading, poker playing machines around.

 

[Highlander]

And you can adjust the multi fine? I couldn't when i tried using that bios.

Yip perfectly ok From the usual 2 places in bios...

Used to run the vDImm mod on the old P4C, but didnt borther with this new board due to not needing 3v+ with 4 dimms anymore....

It seems often to be the case that the hotter chips clock a little better for some reason, not sure why though lol!

 

[Highlander]

Sorry to double post, but topics are totally different

Finally managed to get some decent cooling on the 750 on Monday, got a maze3 mounted (see pic below) It's a bit of a bodge job, but temps are ~ 20-30 deg lower than with the stock cooler, load @ 2.9ghz w/1.62v is around 40 iirc, much better than the 65+ @ 2.7ghz w/1.57v before

Chips now got some room to breathe too, runs 3d stable @ 261*11 and pi's fine at 265*11 Just gotta slowly up the volts and FSB and see how far she'll go, previous owner had 2.86 stable, and ran over 2.9 at some points, so 3ghz benchable might be do able....
mmmmmm sexy

 

[Sjaak]

Nice one

I've tried to compile some of the results, see attachment. I'm not really satisfied as to how it looks, please comment.

 

[Sjaak]

I also put up some pics of my rig up and running, see here: http://www.ocforums.com/showpost.php?p=4075203&postcount=1045

 

[caater]

Very strange! See what it can do FSB wise. My Dothan doesn't like much over 2.7 even there it's a bit flakey. But it can handle a nice FSB as far as 875 chipset goes.
Scott

i tortured mine with high fsb.. vcore @ 1.6v, vagp 1.95v, 120mm panaflo blowing at NB, ram @ 5:4 or 3:2
highest fsb i got past superpi1M: 316 SCREEN
highest fsb sshot: 337

i think with watercooled NB and higher vcore (asus 875/865 boards vcore=NB voltage) i could hit higher fsb.. no practical use though

one interesting thread: X1800XT benching with Dothan platform
dothan + radeon x1800 seems to be a deadly combo

I also put up some pics of my rig up and running, see here: http://www.ocforums.com/showpost.php?p=4075203&postcount=1045

looking nice!
how are the temps compared to just plain inbox fan?
i have currently inbox hsf.. and it's pretty loud, even at the lowest possible setting in bios.. not so loud though as 9600xt here

 

[lowfat]

Good to see water helped your clocks Can't wait to get a replacement pump now

you have a better mounting job that i will use...zipties.

 

[justthinking]

Anyone got tips on how to get Dothan running at 250+ FSB?
My 740 can do 200x13 at 1.375v (bios setting, actual voltage is around 1.3625v)
However, I can't get it to boot up at 250FSB even after lower multiper to 8x and raise voltage to 1.4v

I am using P4P800SE w/ 1010.001 beta bios. Changed RAM setting to 320MHz,
which should be the same as running 200FSB and set RAM to 400MHz.
Am I missing something here?

Do I need more vcore in order to reach the magical 250FSB? Or do I have to mod my motherboard?

 

[Sjaak]

You will need alot more voltage, think of 1.6v+

 

[BrianH2O]

There has been many discussion about this. It seems that the bios applies the FSB before it applies the VCore changes. So, you are trying to apply the 250FSB to stock vcore which does not work well.

What I do is bo0t at 10 x 201 then use SetFSB from the startup folder to boost my FSB to 275. Pretty transparent actually.

 

[Highlander]

You NEED to boot at 201 then up via windows with clockgen. You should be able to boot at a higher fsb though. however there appears to be a "cap" above which you won't be able to boot. The way to find this is kinda simple, raise the FSB in bios til you can't boot
You CAN go higher than this in windows, but for some reason some of the boards max... If you raise the FSB and restart the machine, quite often you won't get a boot the next time round. This is another indocation of the FSB being too high.

The chips REALLY flying now, benching at 261*11 is awsome (though need to drop the FSB and tru with the tighter 11 latency timing) Running #2 on the ORB for 6800LE's atm, number 1 coming this weekend

 

[Highlander]

There has been many discussion about this. It seems that the bios applies the FSB before it applies the VCore changes. So, you are trying to apply the 250FSB to stock vcore which does not work well.

What I do is bo0t at 10 x 201 then use SetFSB from the startup folder to boost my FSB to 275. Pretty transparent actually.

Would agree, will confirm tomorrow when I bench next, have a vCore read point on the board

 

[justthinking]

I guess there really is no way to boot at FSB higher than around 200 then.
Will play with SetFSB a bit this weekend and hopefully can finally achieve 250FSB.

BTW, do I need more vcore to run at higher FSB but same clock speed?
If I need only 1.375v to run at 200 x 13 = 2.6GHz, what vcore can I expect if I want to run the chip at either 250 x 10 = 2.5GHz or 250 x 11 = 275GHz?

So far I have tested 3 different dothan chips. (two retail box 730 and one oem 740)
Maybe my sample size isn't large enough, but so far I can find no pattern or no way to predict just what kind of voltage these chips will need to run at 200FSB.

740 (OEM) does 200 x 13 @ 1.375v bios setting (1.3625v actual)
730 (RETAIL #1) does 200 x 12 at DEFAULT VOLTAGE (bios setting leave at AUTO)
730 (RETAIL #2) does 200 x 12 @ 1.4375v bios setting (1.425v actual)

 

[Jimbob7]

You may need a slight bump in voltage. Be sure to give the mobo as many volts as you can (AGP voltage etc) above 250-260, you really need some chipset voltage, and you will find the higher you get the FSB the more heat the NB is going to produce. I start getting errors in games @ 38c on NB that was in the summer, so temps are ok now.

If i was you, and as I did myself, just give it 1.600 and see what you max stable is, i never looked back from there.

Scott.

 

[justthinking]

Other than cpu itself, is it safe to assume my board/system will function the same as before? I was using a 2.4C northwood running at 270FSB for over 2 yrs before switching to dothan w/ cpu voltage at 1.55v, ram at 2.85v, and agp at 1.60v.

I know my board can handle 270FSB when couping w/ northwood, but don't know if running dothan at 250FSB will have any effect on my system's stability. (assuming cpu is not a factor and my dothan can run at 250FSB w/o any problem)

 

[Jimbob7]

The mobo should find it easier to run the Dothan at high FSB simply because of less power consumption, so yes, as before you should be able to clockgen to 270FSB, providing the CPU itself is willing to go that far etc.

Scott.

 

[justthinking]

I usually test my system's stability by running prime95 torture test and 3dmark2001se looping infinte for 24 hrs each.
My 740 running at [email protected] can pass prime95 24 hrs w/o any problem, but when testing w/ 3dmark2001se,
it resulted in 3dmark2001se quit itself about 10-11 hrs enter the test.
Raise the voltage to 1.3825v and still can't seem to run more than 12-14 hrs.

CPU temp and mb temp shouldn't be an issue as both were staying consistant at 56/38 for the whole time under full load.

Anyone has any suggestions? Do I need even more vcore to get 100% stability?