To crashman

[Guest]

I now have p 3 750e and asus culs2-c as your suggestion,
So far i just go for 952 Mhz. do you have any suggestion to go to i Gz?
tx

 

[Crashman]

Going to 1GHz will result in:
1.) Better stability for the rest of your system as it returns PCI, AGP, and Memory to normal timings.
2.) Ability to use 4x AGP without the registry tweak now required at higher AGP bus speeds.
3.) Less stability of the processor is possible.

I recommend you at least TRY the 1GHz setting, Try 1.90v. If it's stable, try backing off to 1.85v. If not, try 1.95v. Keep an eye on the temps. Remember that increased temps=increased resistance, so having a good cooler may reduce the voltage required for stability. I recommend 3D-Mark 2000 as a testing program for stability.

Suicide is painless...........

 

[Guest]

I try 130:130:43, and 133:133:33 in both 1,9 and 1,95 v.
The computer does not post.

My p3 750 is sl4cf, 128 Mhz pc 133 Apacer.

tx

 

[Crashman]

That's too bad. Of coarse 133/133/33 would be easier thatn 130/130/43 because of the PCI bus, sorry it did not work. There is a method that would probably get you there, but it seems a little extreme. It is called the wire trick. The problem is, that until your compter post, it will be at stock voltage with many motherboards. Then, BIOS changes it during POST. The solution is to modify the voltage detection pins so that the CPU is set at a higher default voltage so that it can actually reach POST.

Suicide is painless...........

 

[peteb]

you can do this with the (L7?) bridges on the chip too. Possibly easier although it will probably require breaking and remaking. 6 of 1 1/2 dozen of the other I guess.

-* This Space For Rent *-
email for application details

 

[Crashman]

Sorry, we are speaking of the PIII, not Athlon, and of voltage, not multiplier adjustments. There are 4 voltage detection pins on the PIII, and base voltage is determined by which are or are not hooked to the fourth.

Suicide is painless...........

 

[peteb]

Ahh - yes um.... I'll get my coat shall I?

Sorry guys, I really thought (for some strange reason) we were talking Duron here.

I take the baord doesn't have jumpers for cpu voltage though? Thant would get round waiting for BIOS to initialise and set it higher.

-* This Space For Rent *-
email for application details

 

[peteb]

but also I think you can change the voltage, not just mult. using AMD bridges. Not relevent here of course though.

-* This Space For Rent *-
email for application details

 

[Guest]

Crashman could you tell me in ease way?
You also have one CUSL2-C, rigth?
what is your cpu, and what speed you have reach?

 

[Crashman]

I have a CUSL2 and a 700E. I have been able to reach 1021 stably, but I turned it back down to 933 to reduce heat and voltage for longevity.
The 700E is actually the best choice for overclocking, as the PIII has a "hard sport" at around 980MHz, where passing this speed requires larger voltage increases, and is not as likely to work.
It takes only 1.75v for me (and most other users) to get to 933 with total stability. It takes 1.85v to go to 980, and 1.95v to get to 1021. So you can see that things are getting pretty tough at around 980MHz. Most processors fail to pass that mark by any significant amount.
The 700 would have given better performance at 933 than your 750 does at 952, do to the increased memory bus. Also, you would have had an easier time with your cards, due to PCI jumping down to 1/4 bus when you reach 933 (and AGP dropping to 1/2 bus).
The real deal is that even at 980, with the 700, you are still running only 35MHz PCI and 70MHz AGP bus speeds, and are still getting true 4x AGP and full speed memory timings without the registry hack. You see, above that AGP speed, your motherboard has a protection system that secretly backs down AGP to 2x and reduces memory timeng to protect your system. There is a registry hack implemented by a program called stuffpcr that returns it to it's normal (full speed) settings.
So you should have got the 700. Maybe somebody will trade you?
Alternatively, you can use the pin trick ("wire trick") to force your processor to be detected at a higher voltage, so that it can reach BIOS and boot.

Suicide is painless...........

 

[Guest]

Is there any spesific stepping like cCO or cBO in your opinion?

 

[Crashman]

Some people prefer cCO stepping, but I haven't seen a big difference between the two, except that it has a slightly higher stock voltage (1.70 instead of 1.65v), which may make it easier to boot without the pin trick.

Suicide is painless...........

 

[Guest]

could you tell me about wire trick?
because many people in pc data base in overclockers.com, overclockers.com.au can go 1 Gig.
and I hardly find p3 700e in my country right now.

 

[Crashman]

There are four voltage detection pins (VID 0, 1, 2, 3) and a common pin (VSS). The voltage detected depends on which ones are connected to VSS. If you have a cBO stepping CPU (1.65v), the easiest thing to do is connect pin vid3 to pin vss, which gives you 2.05v, a little much but many people are using this. The rest are already connected. The next easiest voltage is 1.95v, which requires Vid3 to Vss, and the removal of Vid1.
If your processor is a cCO stepping, you are lucky, because you can get an almost ideal base voltage of 1.80v by connecting Vid1, or 1.90 by connecting Vid2.
So determine which stepping you have (it will be the lowest available voltage in BIOS) and get back to me, and if you send me a private message I can email you the chart and graphic.

Cast not thine pearls before the swine