Help overclocking E1400

Ducky Spud

Ducky Spud

Ducky Spud

Associate
Joined
24 Feb 2009
Posts
1,758
Location
Teesside
I know its a useless processor for games, the L2 cache is tiny etc etc, its hopefully just filling a gap until I can afford somthing a little better.

Anyway, Im not completely new to overclocking and I know this processor is capable of more, I just cant get there and am not sure why! The photos show my current settings. Ive basically increased fsb to 300, upped the vcore (its over 1.5 on this but boots at 3Ghz with about 1.47) and set the "system memory multiplier" manually. Also changed the voltages slightly for the DRAM and MCH (northbridge right?).

P1040643.jpg


P1040647.jpg


Ive also disabled some of the settings in here

P1040646.jpg



Most guides say about setting the RAM divider, guessing thats the same as my memory multiplier? Just worked out differently? I can get mine to set my RAM frequency to just over 900 before it becomes unstable.

Also, Ive read about setting the NB strap, is that "MCH frequency latch" on mine? Ive heard that I should set it as high as I can (over 1000 I was told) but the highest setting for it is 400Mhz at which point my PC wont boot anyway.

Anyway, you can see my settings as they are. If I raise the fsb even by 5Mhz it wont boot, even when I raise my vcore upto around 1.54 (like I said, it boots as it is at about 1.47). This makes me think it must be some other settings? Or have I just reached the limit for this chip? Hope not as Ive seen quite a few running 3.4/3.5Ghz

Thanks for any help :)
 
Hey Ducky Spud,

cool man, sorry to make you work some more but could you copy this layout, it includes a lot of helpful info for us trouble-shooters to work with thanks! :cool:

Off-Topic Sample Help Screenie!
samplehelp.gif


[edit] This screenshot is preferably done just after a *Load CMOS Defaults* as well as half way through an overclock! :o
 
Last edited:
Hey Ducky Spud,

I think you need to note down your BIOS changes and then load Set-up defaults abd start from scratch! :p

With the BIOS on full auto the system should set-up your memory using the 200-Strap and a 1:2 Memory multiplier . . . . i.e FSB will be 200MHz but Memory will be 400MHz/DDR2-800.

Boot to Windows and take another screen like above and have a quick surf, the system should feel extra nippy as the Performance-Level (tRD) should have been auto set to 6 (pls confirm if mem frequency is 400MHz).

Then reboot and change the FSB-Strap/Latch from [auto] to [266] then save n Exit . . . back into BIOS and manually set the vCore to what you feel your chips needs at 2.66GHz then then save n Exit . . . back into BIOS and dial in a 266MHz-FSB then save n Exit and boot to Windows!

You should now be sitting in Windows with a 2.66GHz Processor and the memory running at 400MHz/DDR2-800, if so can you grab another screenie like above and we can proceed to the next step! :)

[notes] If you have the time it would be good to know how your particular system works on full [Auto], in theory if your mobo is playing nice you could do all the above steps by simply dialing in a 266MHz-FSB, the chipset could change the vCore and FSB-Strap by itself . . . can it?
 
Hey :)

Screen of it running all on auto

AllAuto.jpg



And Screen of it running at 2.66Ghz. Just as a note, with my FSB Latch and FSB set to 266MHz, it set my memory to 1066Mhz and failed to boot. I had to manually set my mutiplier to 3.00A (set it back to 800MHz and A because thats for 266FSB... I think).

266fsb.jpg



Same happened when I just changed the FSB to 266, RAM set itself to 1066MHz. Goes through the startup but BSOD's just befor it shows the desktop. Think its just my RAM holding it back doing this.

That all alright? :)
 
Hey Ducky Spud,

The board did something unexpected there! :(

It should have automatically changed the memory to run at (2:3) 400MHz/DDR2-800 when you switched from 200MHz-FSB to 266MHz-FSB :confused:

For some reason it kept the memory running at 1:2 . . . i.e it's not auto changing the strap? :rolleyes:

Can I ask if you adjusted those settings one at a time . . . saving and exiting in between or did you apply them both at once (FSB-MHz + FSB-Strap?).

I'm trying to get your board to act as if it thinks you have a native 266-FSB chip but for some reason its not detecting your memory [SPD] properly.

Is there anyway coming from a full CMOS reset to get your board to automatically select the (2:3) memory divider with you only having to select either 266MHz FSB or 266-Strap?

Anyway leaving that glitch aside for a moment try these settings, I want to see if your Northbridge [NBCC] is happy running a decent frequency (at least higher than 300MHz!) . . .

Intel® Celeron® E1400 (10x200) 2.0GHz

8 x 266 = 2128MHz
NBCC = 332.5MHz

7 x 266 = 1862MHz
NBCC = 380MHz

6 x 266 = 1596MHz
NBCC = 443MHz

Are all your volts still on [auto] except vCore or?

If you can't get any of the above to boot then please methodically apply increased levels of vNB (Northbridge voltage!) until it hopefully works, make sure the Northbridge heatsink isn't getting scorching hot!
 
I did them all together, Ill try doing them 1 at a time and see how it goes. Just at uni at the moment, will try it when I get home in a few hours.

I set everything to auto including vCore, it boots and runs stable at 2.66Ghz on standard voltage. Is the MCH Core the Northbridge voltage in my BIOS? I'll have a go trying those settings when I get home as well.

Thanks for the help again :)
 
Right, I tried setting them to 266 one at a time, but when I raise the FSB the memory frequency just goes up at the same time, can see it changing as I increase the FSB

As for changing the multipliers, tried them, all booted fine and ran stability tests (only for 10mins each), all with vcore and all other voltages still set to auto.

Here is a shot while at 6x Multiplier
6xMulti.jpg


Is this done right? With the FSB still set to 266MHz?

Hope this helps in some way!
 
Right, if you can boot your Intel® Celeron® E1400 at 6 x 266 and the computer POSTs Ok and runs Windows stably thats good as its telling us your able to run 443MHz-NBCC which is basically the same as 443MHz-FSB.

Now looking closely at your SPD tab I just noticed the NANYA memory hasn't got complete JEDEC SPD info, it goes

  • 200
  • 266
  • ---
  • 400
I've not seen that before (missing 333MHz SPD) and dunno if its causing a problem but it's something to note none the less

Ducky Spud said:
If I raise the fsb even by 5Mhz it wont boot, even when I raise my vcore upto around 1.54 (like I said, it boots as it is at about 1.47). This makes me think it must be some other settings? Or have I just reached the limit for this chip? Hope not as Ive seen quite a few running 3.4/3.5GhZ

P1040647.jpg
Click to expand...
  • CPU Vcore = 1.4000v (should be enough for 3.2GHz CPU?)
  • CPU Termination = Auto (This may need manually adjusting)
  • CPU PLL = Auto (This may need manually adjusting)
  • CPU Reference = Auto
  • MCH Core = ??? (could be 1.2000v to 1.4000v??)
  • ICH I/0 = Auto
  • DRAM Voltage = Whatever it's specc'ed at

Ok so having re-read your post above you are saying the best clock you got so far is 10x300= 3.0GHz, if you set 10x305=3.05GHz the PC will not boot! :confused:

That's almost certainly nothing to do with the CPU and *everything* to do with the Northbridge/FSB-Strap! :)

A sick overclocked CPU will still POST even if it fails to load Windows so I'm pretty confident the problem lay elsewhere which is good because now you can turn down that vCore lol! . . . . 1.5000v is waaaay more voltage than that chip needs! :D

I'm having a mul over this, in the meantime you need to try a lot of different variables of FSB, CPU-Multi and FSB-Straps . . . . basically taking a different approach to get the CPU running over 3.0GHz

9x355
8x400
etc

So that's lots of different CPU multi's, FSB-MHz, FSB-Strap and Voltage combos for you to play with, methodically work your way through the few dozen variables hehe! :p

Remember take baby steps with your motherboard and save n exit after you make each adjustment, also don't assume that a particular setting does whats it meant to i.e chaos theory! :o
 
Yea, will try some of these. Never tried lowering the multiplier and increasing the FSB lots, might make a difference.

What about CPU Termination and PLL? Ive no idea what these do at all haha. I'll just nock them up 1 or 2 steps so its not too much and test

Thanks for the help, im off to get some tests in!
 
Ducky Spud said:
Yea, will try some of these. Never tried lowering the multiplier and increasing the FSB lots, might make a difference.
Click to expand...
It may well do, through trial and error you basically build up a better idea of what the problem is!

Ducky Spud said:
What about CPU Termination and PLL? Ive no idea what these do at all haha. I'll just nock them up 1 or 2 steps so its not too much and test
Click to expand...
CPU Termination (FSB Termination/vTT) is a setting I still don't fully understand however it definitely helps to boost this in a situation where you get a No POST similar to what you are experiencing. I've had a few overclocked builds that mostly POST fine and can run Prime for days however once the PC is rebooted it sometimes fails to POST, raising vTT deffo helps here!

I also think there may be a link between vTT and vNB (aka vMCH/Northbridge volts) but I'm testing that myself atm and wont know the answer for a week or two . . . in the meantime I would suggest adjusting these seperately at first and if that doesn't help try adjusting them together so their values are similar!

CPU PLL is again a setting I've not had to use a lot (normally have it 1.52-1.6v) but from what I gather its a handy tweak for helping CPU's that are nearing their max frequencies to keep going a bit more!, you've pushed an extra 1GHz out your little Celeron so it may well be it needs some advanced tweakage!

Big.Wayne said:
  • 200
  • 266
  • ---
  • 400
I've not seen that before (missing 333MHz SPD) and dunno if its causing a problem but it's something to note none the less
Click to expand...
This SPD info from your memory is bugging me, if I had your machine here I would be swapping out that NANYA for some other DDR2 to see how that *interacted* with your motherboard.
 
Well Ive tested a few things. Using 9x multiplier didnt achieve much more. 8x managed to get into windows at 3.15, ran a simple benchmark fine then tested with prime95. core 1 failed after about 10 seconds and BSOD about another 30 seconds after that. Upped the voltages a bit and managed to get it to last about 30 seconds before core 1 failed, so still not much better.

I'll just keep messing round, see if adjusting voltages more will improve the stability.

And I know the RAM isnt the best, but the only other RAM I have is 4 old sticks of DDR, not going to help much I dont think!!

Like I said, will keep messing round, see if I can get it running a bit more stable on a lower multiplier
 
Ok Ducky Spud! :)

The first big step is actually getting the PC to POST, if you can do that at various different configs then its promising.

The second big step is to get the thing to actually load Windows!

Getting stability will follow! :p

Take your time with it, you got quite a lot of variables to work through. If you get in a jumble then Load CMOS defaults and start again.

I will be testing an E5200 Wolfdale in my main machine soon, that also has a 200-Strap and had problems overclocking using anything over 350MHz-FSB in my uATX system, so hopefully I will have some more feedback for you soon! :cool:
 
Yea, when it wasnt even getting to POST it was quite annoying! Got it now on 356fsb and 9x multiplier making it just over 3.2GHz, not the most stable prime95 wise but boots to windows and can cope with most tasks so far.

On prime95 core1 fails after about 5 minutes. Not the best but like you said, got a lot to work with yet! Do you think there is a reson core1 seems to always be failing first? Or will it just be that its just the random factor of some overclocking better than others and for some strange reason my cores being different?

Well thanks for all your help :) Definately doing a bit better now than I was before! Good luck with the E5200! Let me know how it goes :)
 
Ducky Spud said:
Yea, when it wasnt even getting to POST it was quite annoying! Got it now on 356fsb and 9x multiplier making it just over 3.2GHz
Click to expand...
A bit of progress then which is good! :)

Now your running over 333MHz-FSB have you tried manually adjusting the FSB-Strap(or latch) to 333?

I'm interested to see if your mobo knows how to cope with the *missing* 333MHz-Memory [SPD] info.

From a reset BIOS is it possible to get back to where you are now in two moves? i.e by selecting 333MHz-FSB and adjusting vCore only?

Ducky Spud said:
On prime95 core1 fails after about 5 minutes. Not the best but like you said, got a lot to work with yet! Do you think there is a reson core1 seems to always be failing first? Or will it just be that its just the random factor of some overclocking better than others and for some strange reason my cores being different?
Click to expand...
That sounds like either the processor is running too hot *or* it isn't getting enough voltage?

How are you Full load (Prime Small FFTs) core temps btw?

If you try adding more vCore and that doesn't help or in fact makes the CPU error faster than you may have a temp issue?

Let us know, good luck! :)
 
The FSB-strap was set to 333MHz when running at 3.2Ghz, seemed to cope fine. Can get some screens for you later if you like when I get home?

I'll try increasing the vcore as well. When I test with small FFTs it runs at about 51-52C, not too hot to cause crashes. Was running stable at over 60C before I got round to do proper cable managment to improved air flow. Should be getting some of my mates spare watercooling parts soon too which might help a bit :)

Thanks
 
You must log in or register to reply here.
Back
Top Bottom