The_Manual

First of all the Northwood cores are based on a 130nm process whereas the Prescott is based on a 90nm process. This should mean that Prescott's are cooler and more efficient, in actual fact its vice versa.

Thermal Control and the abilities of the processor’s to cope with the temperature.

It is believed that Intel processor’s based on a Prescott core are able to withstand the higher temperatures, unfortunately this is not the case. They may run much hotter normally but they are not able to operate efficiently at higher temperatures compared to the Northwood core.
The average maximum thermal temperature for a Prescott core is actually 67.7C whereas the Northwood core is 72.3C, this shows you that a Northwood can actually run hotter.
You see when you decrease the process size for example 90nm to 65nm the thermal abilities of the CPU are lower, so it needs to run at a lower temperature, this can change depending on how the processors is built, what material is used to build it and how efficiently it was made.

Now as you go to a lower process you can usually fit more transistors onto the CPU, this again will increase the temperature of the CPU because each of those transistors are getting used and are giving out heat contributing to the overall CPU temperature.

Enough of this, now onto thermal control.

Thermal Control: This is technology incorporated into CPU’s to improve their ability to handle temperatures if they are going higher than the CPU is willing to cope with.

There are currently two types of thermal control, Thermal Control 1 (TM1) and Thermal Control 2 (TM2).

TM1: If the CPU reaches a certain temperature this technology inserts idle clock cycles into the system to try and cool it down. This is called Thermal Throttling. As it is inserting idle clock cycles this means that less clock cycles can be used to process data, thus a decrease in processing power is noticed.

TM2: This is the main part of Intel SpeedStep Technology. This can decrease the multiplier on a CPU to x14 and decrease the CPU voltage (vcore), thus lowering temps but reducing the clock speed of the CPU. This is only activated when a small amount of processing power of the CPU is actually required. TM2 is part of the Enhanced Intel SpeedStep Technology that is implemented into the Intel 6xx, 8xx and 9xx processors and obviously the mobile processors.

Pipeline Architecture

The pipeline architecture is very important in relation to the speed of a CPU, this pipeline architecture is the amount of stages to execute an extension or anything relevant.
The Northwood core has a 20 stage pipeline architecture vs. the Prescott’s 31 stages.
The longer the number of stages the higher the possible clock speed of the processor but the less efficient it is as it takes much longer to execute what is needed. The less number of stages that you have the lower the clock speed but there more efficient the CPU is. The main disadvantage however with a longer pipeline is that the IPC count is lower (Instructions Per Clock Cycle), this is why it is best to have a middle length pipeline so you can have acceptable clock speeds, a decent IPC count and a improved overall efficiency.

Branch Predictor
There is something in a CPU called a branch predictor, it predicts what you want to do, however if the predictor makes a mistake it has to go back down the pipeline to trace the mistake that it has made, the longer the pipeline the longer this procedure takes. However the chance of a mistake is less than 20% on Intel system's. The branch predictor on the Prescott is much more sophisticated than the one in the Northwood core.

Cache and FrontSideBus

The latency on the cache in a Prescott CPU is higher than that of a Northwood processor, this is true for L1 and L2 cache, I’m not sure about L3 cache is it is rarely used for desktop processors anymore. The speed of the cache in a Intel Pentium 4 processor is the same as the clock speed of the processor.

Northwood: 512KB L2 cache,
Prescott: 1MB L2 cache,
Prescott 2M: 2MB L2 cache

These values do not include the Intel Celeron D processor line.

Execution Trace Cache

This is a lovely bit of Intel CPU’s, the execution trace cache is part of the L2 cache that stores “decoded-micro operations” so that when an instruction is needed it can get it directly from there instead of having to fetch it from a slower location and decode it. It is much faster this way as the CPU can access it directly and at a lower latency, also the cache will be much faster as it is running at the speed of the CPU. The micro-operations will be put into a predicted order in the execution trace cache so that they can be executed faster if the prediction was correct. The latency of this cache is lower on a Northwood CPU than on a Prescott, however the Prescott has more execution trace cache than the Northwood core.


Now onto the FrontSideBus of these cores.

The Northwood is an older core so it would be logical to guess that the bus speed of it is lower, that is correct to a degree, however, both the Prescott and Northwood cores have 533 and 800MHz FSB speeds so there is not much difference here. The Prescott however can have 1066MHz bus speeds on the Extreme Edition’s.

How to calculate the FSB speed

The FSB is not what you think it is, you may have heard the terms “quad-pumped” but you may not know exactly what it is.
Quad-pumped is a way of creating a faster “theoretical” bus speed. What it does it multiply the actual bus speed by 4 because 4 electrical signals are being passed per clock cycle. Therefore you will actually have a lower bus speed than you actually think, its just having more electrical signals sent down it.

e.g.
1066MHz FSB speed = 266MHz bus speed x 4 (quad pumped architecture)
800MHz FSB speed = 200MHz bus speed x 4 (quad pumped architecture)
533MHz FSB speed = 133MHz bus speed x 4 (quad pumped architecture)

I would have a lot more to write but I am limited by the website on how many words you can put in :D


Last thing then, the REE

Welcome to the world of Netburst, the REE stands for The Intel Rapid Execution Engine. This is incorporated into all the cores based on the Netburst architecture.
You should know what a ALU is (Arithmetic Logic Unit) if not I will tell you a mini bit about it, its just a unit in the CPU to process arithmetic calculations. What the REE does it actually make the ALU’s operate at double the clock frequency of the CPU e.g. 3GHz CPU will have ALU’s running at 6GHz. As the Prescott can have higher clock speeds it will usually have higher speed ALU’s depending on the overall clock speeds. As the clock speeds of the ALU are very high they also contribute the CPU temperature.

I think that will do, hope you enjoyed the read :)

This information may not be copied, modified or used without the permission of me or admin first.
Copyright© Overclock.net 2006. All Right Reserved.

Note: If there are any mistakes etc then give me a PM ;)

thehybridpyro

finally the TECHieish info
thanks this actually is really helpful

__________________
Looking for cheap compressors for phase change