How do I explain to my dad that a 1.5 ghz dual core is not a 3.0 ghz single core?

To him, he thinks that (clock speed) * (# of cores) = "total" clock speed.

So a 2.2 ghz dual core is not a machine that clocks at 4.4 ghz and a 1.8 ghz quad core is not a 7.2 ghz beast.

Hard to explain, but it's faster than single-core. This is probably an inaccurate way of demonstrating it:

Assume something gives 100% power (this would be single-core), now dual core would be more like:

100% + 50%.

So it's like half of what you previously had. So you are gaining more power and speed, it is BETTER than single core, but it's not double the power. Does that make sense? Also, if I am wrong, I would like to be corrected. That is all. I'm sure others here can either correct me, or explain it better. My analogy probably made very little sense as I'm not too experienced in comparing the two.

The hertz is not a performance measurement but, rather, a cycle-frequency measurement (the amount of cycles per second.) The power of each cycle in combination with the amount of cycles is what determines performance. To realize the difference, it would be helpful to imagine that you have two bikes in front of you and they are both trying to peddle themselves out of the mud. One bike has tires without much traction, and the other has tires with a lot of traction. Every rotation of the tires on a bike would represent a cycle of work. Now, imagine that both bikes were spinning their wheels at five cycles per second (five hertz.) The one with the traction will get out of the mud quicker, not because of the amount of cycles, or rotations of the tires, but because more work was done with each cycle. I hope that was clear.

In the processor industry it is not much different. This is why you can place a 2ghz AMD Athlon against a 2ghz Intel Pentium 4 and see the AMD perform better, in the previous example, AMD would be the one with the better traction. In general, it is better to compare performance by actual testing and relate the gigahertz to the consumption of electricity. Each cycle of work requires a certain amount of electricity so the more cycles the more electricity will be used. Consider the amount of energy would be required on the bicycle example earlier, the one spinning his tires endlessly without traction may get out of the mud but it requires more energy.

A dual core processor is sort of like adding more traction to the tires. A 2ghz single-core processor will have less traction than the 2ghz dual-core processor. Both processors will use the same amount of cycles but the one with more traction will get out of the mud quicker. Of course, like one would expect, the issue is a bit more complicated than that, since software needs to be written to take advantage of the extra work per cycle. But, in general, two peices of software will work with one core each so they will both get the power of a whole processor so multi-tasking will improve quite a bit. Again, this is simplified but it proves the point I am trying to make. For software written for multi-core processors you'll notice that they perform much better; software written for single-core processors won't really show a difference; regardless of the software being used, when multiple peices of software are being used one will notice an increase in performance because more tasks are completed per processing cycle.

I hope that clears things up.

For a non-metaphorical answer http://answers.yahoo.com/question/index?...100AAPc70i

Hard to explain, but it's faster than single-core. This is probably an inaccurate way of demonstrating it:

Assume something gives 100% power (this would be single-core), now dual core would be more like:

100% + 50%.

So it's like half of what you previously had. So you are gaining more power and speed, it is BETTER than single core, but it's not double the power. Does that make sense? Also, if I am wrong, I would like to be corrected. That is all. I'm sure others here can either correct me, or explain it better. My analogy probably made very little sense as I'm not too experienced in comparing the two.

This is wrong to some extent.

A 1.5 GHz dual core will at best perform almost as well as a 3.0 GHz single core in some tasks. Almost because it requires some clock cycles to finish parallelism and such.
A 1.5 GHz dual core will at worst perform as well as a 1.5 GHz single core.

Sometimes you can split the work almost equally up, for example when summing numbers. If you have a list of numbers and split the list into two lists, ask the two cores to sum their lists up and then ask one of the cores to sum the sums, it would almost take the same amount of time. (For large lists you would notice no difference in speed.)

Sometimes you cannot split the work at all. Say that you want to calculate the nth fibonacci number. You won't be able to speed the process up because every fibonacci number is dependent of the earlier fibonacci numbers, and since you haven't calculated them yet, you're not able to speed this up. (With the formula F(n) = F(n-1) + F(n - 2))