Author Options:

Just how fast can a 2.6GHz Pentium 4 go? Answered

Well I have gotten a 2.6GHz Pentium 4, on the socket 478 with stock cooling, to go as high as 2.94GHz before the system starts to hang. I believe I have the Core voltage set to 1.875v, and I am running windows XP home edition.
the system specs (other than the cpu) are:
-Ram-1gb in four slots
-HDD's-Fujitsu 10gb and Maxtor 10gb
-Mainboard-Asus P4GBX
-Video-Nvidia Geforce (2?) pci card
-Cooling-Stock pentium 4 cooling aparatus and 2 case fans
-PSU-350W Enermax

The cpu is just a standard socket 478 pentium 4 that has a stock speed of 2.667GHz

Has anyone gotten theirs to go any faster?


64 feet per second, from a third-floor office. Maximum speed without overclocking would be about 200 miles per hour, from a FedEx cargo hold in-flight.

what about air resistance? that would slow it down a bit, as your average desktop tends to be a fairly slab-sided entity. I'm guessing that slightly below 100mph would be the terminal velocity.

Nope, I did a simple calculation, and looked up some results. Terminal velocity for a sphere (which is a reasonable approximation to a desktop cube) is 211 mph.

I assumed the box is tumbling randomly, rather than falling with one face presented preferentially. That randomness makes the sphere not unreasonable, I think. Besides, the symmetry makes the calculation easier :-)

I forgot to root it. 40x40x20 at 10kg was what I went for. shape & area determine drag - what size sphere?


I took a desktop to be about 50 cm average dimension, so a 25 cm radius sphere.

That gives 6.5 litres doesn't it?
Depend upon the case I guess?


I get 65 liters for a 25 cm radius sphere (15625 * 4p/3 / 1000). I used a density of 1.5 times water, to account for the difference between the plastic and metal stuff, vs. the air spaces, and I used a drag coefficient of 0.5. A bit more research shows that I should probably have used something between 0.8 and 1.0, for a cube falling at an arbitrary orientation.

Like I said previously, I forgot to root it, but it seems slow if I do?

Watch video - "approaching 320 m/s", "penetration: 40 metres"


I'm pretty sure 211 mph is too high, as the fastest a person can fall is about 210 mph, if they are trying, and a PC has a lower density than a person while being less aerodynamic.

Um. A PC, especially a modern one, has a substantially higher density than a person (the case is smaller and has less airspace, and the materials inside are all denser than water). Terminal velocity for a "tucked" skydiver, or a freely dropping bullet is about 200 mph.

You can argue over * or - a mile per hour or two, but it's still gonna hurt with the sudden stop.

At least the sudden stop from 200 ish mph will be more spectacular than just pushing the processor too far, and letting the smoke out.

The volume of the average PC is about 23,300 cm3

so the equivalent volume of human has a mass a bit less than 23.3 Kg.

Apparently, the average mass of a PC is between 15 to 30 Kg, depending on the specific case and hardware. I would have said the PC in question is somewhere in the middle of this, so it would probably come down to how much more aerodynamic a human is than a PC, and the altitude.

"I believe I have the Core voltage set to 1.875v" - worries me, as you should know.
How are you over-clocking it?


the mainboards bios lets me adjust the core voltage and frequency if i set it to manual.

i either have it at 1.875 or lower, any higher and the system hangs.

Do you follow a particular method / procedure? Fiddling with core voltage is something best left unless a person really knows what they're doing - CPUs can die prematurely etc.
Anyway you've squeezed 10% out of it which is something.
(Anyone else)


i think if i want it to go faster i would need better cooling on the cpu, it idles around 60 degrees.

Idles at 60oC? Take the cooler off clean everything and apply a very-small amount of thermal-compound - doing this could easily knock 20 degrees off (unless your cooler is really awful)


its the same small stock cooler that was on the board when i got it, i cleaned it out before i started putting it together, i think its just a bad cooler, but i cant afford a better one, they just cost way too much for my budget.

The seating on the CPU may be poor, honestly - redoing that alone (carefully) can make a big difference. There's often too much thermal-paste


theres not too much, maybe not enough... i did use petroleum jelly instead of any thermal paste, I cant afford all of that fancy thermal paste. for a while i "recycled" it from an old stereo.

I think a better cooling system is in order.

Well there you go "i did use petroleum jelly". Fix that first.


who thinks i should try this with a socket 7 cpu? (amd k6-2 or pentium 1, maybe a newer celeron?)

.  For maximum practical speed, you'll probably be better off asking on a forum for over-clocking.
.  The two main limits I can think of are heat and capacitance. Because of the way most chips are made, you can only take so much heat out before physical stresses become too great. Even if you could get a liquid Hydrogen bath to work, capacitance would still limit how fast you could go. Then there's a limit to how fast silicon, et al, can turn on and off.
.  As cheap as computers are nowadays, if I needed more speed, I'd buy a bare-bones, multi-core system and move my HDD, CD/DVD, &c to it.
.  More RAM nearly always helps and nearly always gives you the biggest bang for your buck. Bump it up to the maximum.