Pentium M Vcore question

[Wasmachineman_NL]

Does anyone know the maximum safe Vcore of the Pentium M? I keep seeing people from ~2006-2007 pushing the Pentium M's to 2.8 GHz with 1.6V but I don't know if it's long term safe for the chip.

Imagine a 765 overclocked to 2.8 GHz in a Precision M70, holy shit. I bet it'd rival a Core Duo in singlethreaded stuff!

Or slightly saner: a 755 at 20x133! Maybe I should buy some 755's, bin them and find a chip that can do 2.66 under 1.5v or so.

[Calbris]

Banias or Dothan? Either way, you absolutely shouldn't run them above 1.75 V or 1.6 V respectively. That's the maximum voltage which Intel has stated in their datasheets, but whether if it is 'safe' depends on your definition.

In practice, as long as you don't exceed these voltages, you should be fine. Plus, I don't see a possibility of you pushing any of these processors near that voltage range, as the machine that you speak of is equipped with a small heatsink. Anyways, you shouldn't be expecting excellent overclocks on a laptop that doesn't include a cooling system with sufficient thermal mass and fans. (look at the Clevo P570WM's cooling assembly for reference)

If you want some pointers, I would suggest you to not exceed 1.58 V with a Banias. Anything higher than that would produce an extreme amount of heat in an average laptop (85~90 C+), especially if it's running at 2 GHz or above. Realistically speaking, not a lot of my Banias SL6N5s were capable of booting at 2.26 GHz, so you should be looking at a stable 1.86 GHz at the very least. If you're lucky, you can expect a stable 2 GHz from a Banias at 1.564 V.

All of these experiments was done on a Pentium M-based IBM ThinkPad R40 with a Silicon Labs CY28346ZI-2 phase-locked loop. Voltage adjustment was done via the old tiny-wire-in-the-socket method, and the front-side bus adjustment was done by bridging pin 55 to pin 54 with solder.

For Dothan, it is slightly better in the overclocking department. If I can recall, the highest overclock that can be done without much difficulty, is a good 2.4 GHz. I do not know if it is stable, though. I did not mess around with Dothan that much, so you might have to find it out for yourself.

Expecting 2.66 GHz straight-out from a Dothan might be a little too optimistic, as the voltage required for a stable 2.4 GHz is somewhere around the range of 1.436~1.484 V. This is already pushing the limits of most Pentium M machines, as going above 1.484 V is asking for load temperatures in the 95 C range. Unless you have a 17" with 2 or 3 fans and a large heatsink, don't bother.

Note: My definition of 'stable' is 24 hours of running Prime95's Small FFTs.

[Wasmachineman_NL]

[Wasmachineman_NL]

[Calbris]

You got lucky, I'll give you that. And don't forget that you're running a C0 stepping Dothan, which was not what I used to find out my stable core voltage range.

[Wasmachineman_NL]

[Wasmachineman_NL]

I'd imagine the 65nm stepping/SLJ8Q would put the V/F curve even lower, I have a Core 2 Duo T7600 running at 1.05V.

But then again that CPU is two generations newer and on Intel's 65nm process.