Comparing processor speeds

I don't keep up to date with the latest processor speeds. I am about to buy a new mac book with the 2.2GHz Intel Core 2 Duo, but i'm not sure of what it is equivalent to. Can someone refer me to a websight or tell me how the core 2 duo compares to other processors.

Thanks in advance.

The measure that most people have an intuitive feel for is the good old Pentium IV. Most people have used one at some point. In general, a Core 2 Duo is about 1.4X faster, clock for clock, than an equivalent speed PIV. So, your 2.2 GHz C2D will be roughly equivalent to a (2.2 * 1.4 = ) 3.08 GHz PIV.

BTW, a pretty similar measure applies to the PowerPC G5, which means that your C2D will be pretty much equivalent to a same speed PowerPC G5 chip.

mac57 said:

The measure that most people have an intuitive feel for is the good old Pentium IV. Most people have used one at some point. In general, a Core 2 Duo is about 1.4X faster, clock for clock, than an equivalent speed PIV. So, your 2.2 GHz C2D will be roughly equivalent to a (2.2 * 1.4 = ) 3.08 GHz PIV.

Click to expand...

I would say it's way beyond that. A Core 2 Duo is at least 50% faster than a Pentium D 3.2ghz, in almost any benchmark.

For a Pentium 4 to get those kinds of scores, it'd have to run at least twice as fast, AND have a faster bus and memory. Although its impossible to know how a P4 would scale up at these kinds of speeds, you'd probably need a 6.5ghz P4 to match a 2.2ghz C2D

The clock speed is not faster than the newest P4s, so you can not really compare the two. Intel realized that clock speeds do not really matter after a certain point. This is after their giant marketing campaigns claiming otherwise.

Zoolook said:

I would say it's way beyond that. A Core 2 Duo is at least 50% faster than a Pentium D 3.2ghz, in almost any benchmark.

For a Pentium 4 to get those kinds of scores, it'd have to run at least twice as fast, AND have a faster bus and memory. Although its impossible to know how a P4 would scale up at these kinds of speeds, you'd probably need a 6.5ghz P4 to match a 2.2ghz C2D

Click to expand...

Zoolook, if I look at the benchmark you linked, it would appear to confirm what I said. Looking at the 3D Mark 06, CPU Test, I would expect the Pentium D 960, at 3.6 GHz to be roughly equivalent to a Core 2 Duo at 3.6/1.4 = 2.57 GHz. There is no such beast of course - the closest I can get on these benchmarks is the 2.4 GHz C2D. When I look at the scores between them, they are pretty close, thus more or less upholding the 1.4x generalization I made. Bus and memory speeds likely account for the remaining deltas.

I originally got this generalization from an Intel site, so I am fairly sure of the basic idea...

...meantime, I would love to get ANY model of Mac with the Core 2 Extreme X6800 shown in the benchmark in it - I didn't know that Intel had made a C2D that fast! That would be one fast Mac! ...perhaps pretty hot though - high speed usually equals high heat.

mac57 said:

Zoolook, if I look at the benchmark you linked, it would appear to confirm what I said. Looking at the 3D Mark 06, CPU Test, I would expect the Pentium D 960, at 3.6 GHz to be roughly equivalent to a Core 2 Duo at 3.6/1.4 = 2.57 GHz. There is no such beast of course - the closest I can get on these benchmarks is the 2.4 GHz C2D. When I look at the scores between them, they are pretty close, thus more or less upholding the 1.4x generalization I made. Bus and memory speeds likely account for the remaining deltas.

I originally got this generalization from an Intel site, so I am fairly sure of the basic idea...

...meantime, I would love to get ANY model of Mac with the Core 2 Extreme X6800 shown in the benchmark in it - I didn't know that Intel had made a C2D that fast! That would be one fast Mac! ...perhaps pretty hot though - high speed usually equals high heat.

Click to expand...

I agree with that, but you originally said Pentium 4, not Pentium D. The Pentium D was dual core of course, whereas a Pentium 4 was single core. I thought you meant the classic P4, not the dual core version, apologies if that was my mistake, but I thought it was worth clarifying.

Yeah, my bad. Thanks for pointing that out Zoolook. I should have said "core-for-core", since my reference was to a single core in each case.

However, this raises the troubling notion that a dual core machine is somehow faster than a single core machine. I have always had trouble with such references that companies make (Apple too) about a dual core machine being "twice as fast" as a single core machine. In general, this just isn't true. Such a machine may feel more responsive because it can do more things at the same time, but in general, when one looks at any one thing it is doing, that thing isn't any faster just because there are more cores.

It requires excellent software design to extract maximum benefit from multiple cores. All too often on a multi core machine you will see a highly CPU intensive, but single threaded, program either occupy one core to 100% while the other one loafs, or swap back and forth between the two cores. The net result is an overall load of about 50% on the machine. More importantly, subjectively the job being performed doesn't get done any faster. In fact, with dual core swapping, it may even get done a little slower. This is a case where more raw GHz per core would be more effective than more summative GHz overall.

This observation continues to motivate my desire for faster and faster single cores. The fact that you can now get lots of them in a single package is good, but I want each one to be faster. Try doing any heavy handed work in Photoshop and you will immediately understand why this is of value!!

Thanks guys that has helped me out alot.

Just out of curiosity, which processor has the fewest pipelines ??

Nagger Pumper said:

Thanks guys that has helped me out alot.

Just out of curiosity, which processor has the fewest pipelines ??

Click to expand...

The PowerPc...

If you're only interested in intel, the P4 had 20 pipeline stages, the CD has 12 and the C2D has 14. The C2D has 4 decoders though, with the CD having only 3. 4:14 is a better ratio than 3:12 (3.5 vs 3) which accounts for the C2D's 17% or so benchmark advantage.

Mac57 - I agree with you. Increasing processor speed or adding cores doesn't result in linear system speed increases. There is still the same amount of RAM, still the same system bus feed, plus you have the overhead of the OS having to distribute the workload evenly between cores/processors.

From my overclocking days, th best results always came from overclocking the FSB speed, not from merely upping the CPU speed. Obviously this is harder to do, because there are more components that can give up when the motherboard is put under strain, but often CPU speed increases do little to improve things.

It's more about efficiency and throughput now, rather than clock speeds. A chip with a FSB half the size of a chip with a FSB of 1600mb, will far under preform, even at a much higher clock speed. Besides, clock speeds just tell the cycles per time vs data processed per amount of time.

Village Idiot said:

It's more about efficiency and throughput now, rather than clock speeds. A chip with a FSB half the size of a chip with a FSB of 1600mb, will far under preform, even at a much higher clock speed. Besides, clock speeds just tell the cycles per time vs data processed per amount of time.

Click to expand...

That is certainly true of processor speeds, but not really FSB speeds. The quicker data can get from RAM to CPU to GPU etc, the quicker everything will run.

Mhz vs architecture only really applies to the CPU speed, not the motherboard speed.

My processor speed is bigger than your processor speed...