Pentium III @ 1.95V Safe?

[Peridium]

Anybody here have a Pentium III running at a core voltage of 1.95V? I'm considering if it's worth it to overclock to 1GHz.

If you have a PIII running at high voltages please post a reply with your specifications and how long you've been running at that voltage.

Thanks!

 

[Peridium]

Nobody feels like responding?

 

[ninjazed]

Sounds pretty hot to me. I wouldn't risk frying the chip with that kind of voltage to get what might be a minimal performance boost. If you've got the cash to burn though, fry away!

 

[Fingers]

mine is running at 2.3 just fine, although its a celeron (1). I wouldn't push the voltage that high. most people 1.9 is pushing it. if you do have it running that high i assume you'll have a dead cpu in a year or less. whats it running at right now, most likely you won't notice any difference.

 

[Wiz]

Just make sure you have very good cooling.
There are people out there wiring pins together and running p3's at over 2V
You have to decide for yourself if you can afford the loss of the chip, do your own risk assessment.
Also, decide what the chip lifetime should be. If you plan to upgrade to a newer cpu within a year anyway and that kind of thing. No one can tell you how long your chip will last if you push it, maybe a second, maybe a year and maybe longer.

 

[Peridium]

Well, if you haven't looked at my specifications yet, I'm running my PIII 700E @ 966MHz. Cooling with an Alpha PEP66T and Arctic Silver grease.

Right now I'm using a core voltage of 1.8V and the max temperature I get using MBM 5.03 is 41°C. I've tried running at 1GHz with 1.9V but it wasn't stable in games. The max temperature at 1GHz was around 43°C while running Prime95 or RC5.

 

[Spoooon]

Why don't you head on over to Arctic Silver and follow the guide to calibrate the internal diode on the cpu. You might be surprised by how many degrees your motherboard is off by. If you do take the time to calibrate your internal diode, please PM me with your calibration result.

 

[Mikewarrior2]

Spoon,

My test chip is -3C, my personal chip is -5C.
Edit: These are my correction numbers

Peridium,

What MB are you using?


Mike

 

[Peridium]

I'm using an Asus CUSL2 Mainboard by the way.

Calibrate the internal diode? I didn't know you could or had to. Doesn't MBM 5.03 have that CUSL2 option that compensates temperature readings?

 

[Mikewarrior2]

Calibrating the diode is different: When you calibrate you get exact cpu core temps.

However, the temps are still better than external readings even with an un-calibrated diode.

P3-Xeon chips come with a pre-calibrated Diode.


Mike

 

[Peridium]

So in order to calibrate my internal diode I have to connect an external thermistor to my heatsink to measure the temperature? Can I just use the mainboard temperature sensor for that?

 

[Mikewarrior2]

Yeah, you'd need another thermsitor or thermometer. MB temps vary from Ambient Case temp depending on your case cooling.

Plus, you do want to measure the temp of the air going into the heatsink/fan.


Mike

 

[Spoooon]

Thanks for the reply Mikewarrior. If you happen to read this thread again, do you know if the correction for all CUSL2 mobos is the same? If it isn't, shouldn't it be?

 

[Mikewarrior2]

no, it shouldn't be. Each CPU internal diode is different, therefore each individual cpu must be calibrated to its slotket/mb combination.


Mike

 

[Spoooon]

Shouldn't Intel be using diodes of the same rating for every processor? Granted, you would have to take into account any kind of tolerances present in each diode, but other than that they should perform near the same. Though, I only have a basic understanding of how all this works.

 

[Peridium]

Ok, so my diode readout may not be 100%. I took the time to lower my clock speed to 466MHz (7x66) to see how much of a temperature difference it would make. I kept the core voltage constant at 1.8V. At 466MHz the maximum temperature while running RC5 was 36°C. The case temperature was 26°C. For reference, at 966MHz the max temperature was 41°C with a case temperature of 26°C. Based on these observations, I would estimate that my internal diode is fairly accurate with the temperature readings.

Now back to my original question, how much would the higher voltage affect my processor? Is 1.95V too extreme? Does it not matter if I have good cooling and I keep the temperature below 50°C?

 

[bonkers325]

from my point of view

1.9v is pushing it :Q

1.85v is quite the max i would use :Q

1.95v is just insane

 

[Peridium]

bonkers325,

What voltage are you currently running at? What type of cooling are you using?

Does voltage have more impact on the lifetime of a CPU than temperature?

 

[Mikewarrior2]

Peridium, redo a test at 933/1.8V with the same stressing program.

Looks like your "calibration" is going to around -4/5C.


Spoon,

it should be, but Intel themselves state a +-5C variation. part of the variation comes in with how the individual motherboard(and slotket if applicatble) also read the diode.


Mike

 

[martyt]

Bonkers,

What type of cooling do you use to reach the high PIII numbers? I am trying to get a PIII 700 to its max. I can run at 875, 896 and 910 with various degrees of stability. I would like to run at least at 896 because 875 doesn't support 133 MHz memory on my motherboard.

I am pretty sure that better cooling would help. I just changed my fan from a generic orb to the stock PIII cooler. That alone allowed me to reach 875. Then I bought some Radio Shack thermal paste. Now I can get the machine to boot at 896 and 910!! It will even boot at 933 but is VERY unstable at that speed.

Thanks!

 

[Razzer]

I have a P3 800 with Vcore of 2.05 at 1080 and it has been running for three mouths. I do have a 72W Peltier. Temps are 1c at idle and 32c at full load.

 

[Fingers]

just a reminder that voltage can kill a cpu even if it is at -20C if the chip isn't designed to handle all that current.

 

[Peridium]

Mikewarrior2,

I've clocked back down to 933MHz @ 1.8V and recorded a maximum temperature of 40°C while running RC5. This is also with a case temperature of 26°C.

MBM actually reports the CPU speed as 938MHz though. This is probably because the CUSL2 uses 134MHz FSB rather than 133MHz.

 

[Mikewarrior2]

okay, yeah, your calibration is a -5C. if you use MBM, use the calibration in teh temps catagory, and drop the cpu temp -5C. Then you'll be reading relative results and have the fairly precise actual temp your core is running at.


Mike

 

[Peridium]

Alright, thanks for you help Mikewarrior2.

Now the question is whether anybody else has their Pentium III running at 1.95V with standard air cooling. I'd like to know if anybody has had any troubles with their computers at such a high voltage over extended periods of time.