Maximum Cpu Overclock

[IlijaPlayerP]

Hi. I wanted to know what stops the CPU to go more if the cooling is not a problem?

Let's say I have LN2 Cooling on the CPU, on the Motherboard and liquid cooling for the RAM. The RAM Frequency is not a factor since the ratio can be lowered. What actually is holding the CPU to reach a higher freq.?

 

I've seen two guys with two different mobos and other hardware but the same LN2 cooled CPU and they reach exactly the same frequency.

 

I have two opinions about why the CPU won't go over that frequency:

 

1. The motherboard's BUS frequency is too high for the mobo so it's can't handle more.

 

2. The components inside the CPU (many transistors and other stuff) can't transfer heat fast enough to the cooling area because they're producing heat too fast, so the inside components are hot even if the CPU temp is -150 degrees Celsius (for example) so the CPU can't go more.

 

Is any of my opinion correct, and if not, then what?

 

P.S. And if you know about what holds RAM to go faster you can tell me too.

 

Thanks

Edited December 7, 2008 by IlijaPlayerP

[maj0rgamer]

cpu voltage should play a big factor in stability at higher frequencies as well if i'm not mistaken.

 

my 5000+ be is fine up to 3.0 ghz at stock voltage, but at 3.2 it's unstable and the temps are practically the same even from the stock clock speed in my setup. I have a feeling that if i could figure out how to up the voltage (if my mobo allows it and it should but i havn't figured it out yet) that it would be fine even past 3.2 ghz.

 

you may wanna toy with the voltage if your system is unstable, but knowing what you're doing is important.

Edited December 7, 2008 by Maj0r Gamer

[chuck4456]

No two CPUs are the same.

My Q6600 would run 3825mhz stable (with air) yet my buddies top ended at 3400mhz.

Both GO Steppings - both purchased within the same week.

You just never know.

[Rodneyho]

Hi. I wanted to know what stops the CPU to go more if the cooling is not a problem?

Let's say I have LN2 Cooling on the CPU, on the Motherboard and liquid cooling for the RAM. The RAM Frequency is not a factor since the ratio can be lowered. What actually is holding the CPU to reach a higher freq.?

 

I've seen two guys with two different mobos and other hardware but the same LN2 cooled CPU and they reach exactly the same frequency.

 

I have two opinions about why the CPU won't go over that frequency:

 

1. The motherboard's BUS frequency is too high for the mobo so it's can't handle more.

 

2. The components inside the CPU (many transistors and other stuff) can't transfer heat fast enough to the cooling area because they're producing heat too fast, so the inside components are hot even if the CPU temp is -150 degrees Celsius (for example) so the CPU can't go more.

 

Is any of my opinion correct, and if not, then what?

 

P.S. And if you know about what holds RAM to go faster you can tell me too.

 

Thanks

 

Its the physical limitation of the the hardware at some point. Be it the chip, mobo, ram etc. Lots of time its the chipset (north bridge) that slows things down, thus why the new I7 is going to open a lot up for overclocking. The memory controller is removed from the north bridge and placed directly on the processor. Thus now the memory has a direct route to the processor. Now a LOT of traffic lanes are opened up on the north bridge and I think in the long term that is going to lead to a much better overall ability to OC at higher levels.

[Waco]

Lots of time its the chipset (north bridge) that slows things down, thus why the new I7 is going to open a lot up for overclocking. The memory controller is removed from the north bridge and placed directly on the processor. Thus now the memory has a direct route to the processor. Now a LOT of traffic lanes are opened up on the north bridge and I think in the long term that is going to lead to a much better overall ability to OC at higher levels.

Technically having an IMC makes it harder to OC because you have yet another thing to limit your upper clock frequency.

 

As for general limits, you get to a point where the time to propagate a signal around the die gets high enough to cause errors. This is mostly limited by the overall die size and the minimum feature size. This is why you can get a P4 to absurdly high frequencies, it has built-in delays to help manage signal propagation and the die itself is fairly small.

 

Of course, a highly-clocked P4 is still slow clock-for-clock, but they do hit some incredibly high speeds.

[IlijaPlayerP]

No two CPUs are the same.

My Q6600 would run 3825mhz stable (with air) yet my buddies top ended at 3400mhz.

Both GO Steppings - both purchased within the same week.

You just never know.

 

OK, I wanted to know about extremely high frequencies. Your buddy's CPU is the same as yours but you maybe have different cooler or he has a lower quality motherboard, depends on many factors.

[IlijaPlayerP]

Its the physical limitation of the the hardware at some point. Be it the chip, mobo, ram etc. Lots of time its the chipset (north bridge) that slows things down, thus why the new I7 is going to open a lot up for overclocking. The memory controller is removed from the north bridge and placed directly on the processor. Thus now the memory has a direct route to the processor. Now a LOT of traffic lanes are opened up on the north bridge and I think in the long term that is going to lead to a much better overall ability to OC at higher levels.

 

 

It's the physical limitation of the hardware at some point. - I tought that too but I wanted to know what actually is that limitation but I don't think that anyone will know for sure.

 

The Core i7 is new and I haven't yet put my hands on one, but I prefer P4s (and Celerons maybe) because they go abovo 8GHz

[IlijaPlayerP]

Of course, a highly-clocked P4 is still slow clock-for-clock, but they do hit some incredibly high speeds.

 

 

INDEED

[IlijaPlayerP]

So noone knows what is the limitation of the CPU to go more?

 

P.S. What is the maximum frequency ever reached with a Intel CPU? 8180.4 MHz?