2.4B posts at 166x18 But freezes right after memory count?

[imported_Sigma]

I tried overclocking my 2.4B to 166X18.

In my bios I set the Mem ratio to 1:1. I have 1 module thats PC2700 and one thats 3200 so this is well within specs for both modules. PCI/AGP is locked.

The bios was reporting my Vcore as 1.48 so I upped it to the default of the chip which is 1.525

I rebooted at stock speeds to make sure the above settings took. Everything booted fine and when I got into windows I checked the voltage with my hardware monitoring software it still showed 1.48.

So I went back into bios and checked the Voltage and in the PC Health tab it showed it as 1.52. So I figured no big deal the software must be reading my winbond chip wrong. I went ahead and upped the FSB straight to 166.

After exiting bios the computer gave me a post beep. Then showed the correct readout of P4 2.98 166*18. The computer then showed the Mem speed as 333. Then showed the mem count correctly but right at the next point where it initializes the cards I lost signal to my monitor and the PC froze.

I turned off the Switch and tried again. Same thing. So I unplugged took the cover off and reset the CMOS. Everything booted fine.

-So my question is what does this mean should I have set the Vcore higher?

-Is my chip incapable of even approaching these speeds? I've overclocked before where I would get in windows and then have Prime95 be unstable. Usually bumping the Vcore up fixes this.

-I've never had a system completely lock down like that. I usually dont get to aggressive in my OCs.

-I'm a little afraid of damaging my hardware now. Should I try for a smaller overclock? Or just not bother.

I'm basically just looking for about 10 more FPS to smooth out UT2004 in the rough spots.

Sorry for such a long post thanks to all who take the time to respond.

System specs:

Soyo P4I845PE-lite
P4 2.4B SL6EF
1x512 Crucial PC2700
1X512 Crucial PC3200
Radeon 9500 Pro.

 

[o1die]

I set my 2.4b at 160 fsb with 400 for the memory in dual channel mode on an 865pe board. It's plenty fast. I could get windows to load at 3.0 (167) but it would crash soon after. My cpu voltage is 1.54.

 

[THUGSROOK]

use only 1 stick of memory

 

[imported_Sigma]

Hi Thugsrook,

Why do you say that? I'm not running them dual channel or anything. There both well within specs and I'm not using aggressive timings. Do overclock systems only work well with one Dimm? So you think it was a memory issue?

 

[THUGSROOK]

try it and see if it works.

if not ~ youll need to do more testing to see how far it CAN go, raising it only 5-10fsb at a time

 

[imported_Sigma]

Thanks for the fast reply Thugsrook,

I hope it isnt the memory because using only 512 is not an option for me. I probably did get a little hasty by trying to just grab 600mhz at one jump. I just read at overclockers and it seemed like most people were getting 3.0 ghz.

Say I can only get the FSB up to around 150. 150x18 2.7ghz. 1:1 Mem. Would it be worth it?

Right now I'm running at 133x18. 2.4Ghz but 1.25:1 ratio. So I would be giving up 16Mhz on the memory.

So would gaining 300 more clock cycles have more of an affect than the 16 Mhz memory decrease.

A. 2.4Ghz 333 memory.

B. 2.7Ghz 300 memory.

Which would you choose?

 

[THUGSROOK]

first you need to see what the cpu can do (in 1:1). then work on the memory.
if all the cpu can do is 150fsb then you may as well keep the 1.25 ratio enabled.

it all takes time, there are no shortcuts

g'luck

 

[o1die]

I've had limited luck running two different speeds of ddram. 2 different brands of memory can also cause problems.

 

[Zap]

Originally posted by: Sigma
A. 2.4Ghz 333 memory.

B. 2.7Ghz 300 memory.

Which would you choose?

I would probably choose B.

Like Thugsrook said, "there are no shortcuts." Problem #1 is that you read "most people were getting 3.0GHz." You need to look at the spread. You'll see a whole bunch of people probably starting around 2.7GHz on up. Also, most of the people with higher clocks have specific motherboards (and sometimes extreme cooling and voltage). Your 2.4 -to- 3.0 jump was WAY too high to start with. Start at something mild, such as 140FSB, and go up in 5MHz or so increments. All the collected forum intelligence and experience will do you absolutely no good until you try increasing in increments. Also, try the suggestion of using only one module. This is just to rule out any possibilities. "using only 512 is not an option for me" does not mean you can't spend a few minutes on your overclocking with only 512MB.

Well, enough of my reprimands, now for something useful...

Test with only the PC3200 module in there. Raise core voltage to at least 1.6-1.65v. Keep memory at 1:1 for now. Make sure memory is set to "Auto" for various latencies. Raise memory voltage to 2.6-2.7v. Make sure AGP/PCI are locked at 66/33MHz. Besides working your way up in speed, if you can't force yourself to take the time and just want to go for the gold, try 165MHz FSB instead of 166MHz. I've encountered boards that would work fine at up to 165MHz and would not POST at all at 166MHz - some weird BIOS glitch. My MSI PT880 board was like that with a P4 2.4A until a BIOS update.

 

[imported_Sigma]

Thanks Zap,

I built up enough courage today to give it another try.

I upped the voltage to 1.550.
I set the mem ratio to 1:1
Then I set the FSB to 150.

PC posted, showed the correct P4 2700 18x150
Memory @ 300

Made it to the desktop nothing seemed out of the ordinary. So I went ahead and fired up Q3 to run a timedemo to see if it would be alright. The timedemo ran fine. I was thinking cool now I'm gonna try some Prime95. Just as I was about to exit out of Q3 the the comp rebooted.

So I went back into the bios and set everything back to where I had it. The computer posted but hung on the way into Windows. Gave me a message about Windows could not start normally. I just picked start normally and then it went into windows fine.

So what do you think it is. Should I just up the voltage to 1.6 and go for 150 FSB again?
Did I almost corrupt Windows?

 

[Zap]

Originally posted by: Sigma
So what do you think it is. Should I just up the voltage to 1.6 and go for 150 FSB again?
Did I almost corrupt Windows?

You may have almost corrupted Windows. I've done that before. Just the cost of overclocking.

I personally would crank up the voltage. Nothing to risk except your CPU and Windows, right? Then again, I have enough spare parts and computers that I wouldn't be hurting if I killed Windows or the CPU. If you aren't comfortable with the risk, then maybe be a bit conservative in your settings. Also, even CPU and RAM willing if your motherboard isn't up to the task then nothing you can do about it. I built my GF a computer three months ago with a P4 2.6C that was known to OC fine on an Abit IS7 board. I used an MSI 865G chipset mATX board and I couldn't get diddly-squat of an overclock. I used some OCZ PC3200 and also tried some Kingston HyperX PC4000 RAM. I tried lowering RAM multipliers to 3/2. Nothing. Full stability at default CPU speed, but even 1-2MHz FSB bump would cause glitches. *SIGH* If only some company would make a good overclockable mATX P4 board.

 

[otispunkmeyer]

erm agp and pci speeds will increase as u increase FSB also.......ur graphics card and other pci cards may not like the extra speed on the AGP bus thus games crashing