Thank you for being a valued part of the CNET community. As of December 1, 2020, the forums are in read-only format. In early 2021, CNET Forums will no longer be available. We are grateful for the participation and advice you have provided to one another over the years.

Thanks,

CNET Support

Resolved Question

How fast is my CPU?

Jan 5, 2012 8:14PM PST

In the early days where we had only one core per CPU it was easy to get an idea of the speed. For example a 3.0GHz has exactly that speed.
But these CPUs with cores baffle me. I now have an Intel Core i5-2410M @ 2.30GHz (4 CPUs). What's the speed of my CPU compared to a single core one?
Thanks in advance.

Discussion is locked

thamsen has chosen the best answer to their question. View answer

Best Answer

- Collapse -
It's not even that simple
Jan 5, 2012 9:30PM PST

It's not even that simple, because back in the single core days, AMD was taking Intel to task with their Athlon line compared to the Pentium 4, because for every clock cycle AMD's CPU processed more data, so for many tasks it was actually faster even at a lower clock speed. Which is why Intel kind of went back to the drawing board with the Core line.

But the point of multi-core systems isn't really so much speed as fluidity. Even if one core is busy performing some complicated task, you have another core that's free to handle the routine tasks that make an operating system work, and so you don't notice any reduced responsiveness to the system as a whole. That's the ideal scenario, multi-core setups take a few shortcuts to reduce price from multi-CPU setups, but it is still better than a single core.

So while maybe not a perfect analogy, it should get the basic point across. Think of your CPU like a freeway. The clock speed would be the speed limit, and then the number of cores would be how many lanes of traffic there are. So, say you have 100 cars that need to get from Point A to Point B. With a single core CPU, it'd be like a one lane highway, every car has to travel single file, and so the idea was just to increase the speed limit as much as possible to get each individual car through as quickly as possible. You could liken AMD's more work per clock cycle to making sure that the road is nice and smooth, no potholes or anything else that may slow cars down regardless of the speed limit. Anyway, now if you have 2, 4, 6, even 8 lanes of traffic, all with the same speed limit... In an ideal scenario you can have that many cars moving between Point A and B at a time. They are all moving at the same speed, you just have more of them going at a time.

It's not quite that simple, because multi-core CPUs are such that every core shares the same set of data pathways in and out, so only one core can either send data out or get new data to process. So, despite the best efforts of the CPU engineers who will work very hard to try and make sure the CPU schedules tasks internally to try and avoid this, at some point one or more cores is going to have to sit and wait for another core to finish sending/receiving data before it can send/receive data. So I suppose you can think of it like those really short merge areas anyone who's lived in a big city has undoubtedly seen, where two freeways come together and traffic tends to back up because you might have 2-4 lanes that all converge into one within maybe a quarter mile, and people have to sort themselves out to avoid an accident.

Also, just to correct Mark quickly, each core is running at 3.2GHz. You don't divide the rated speed by the number of cores.

- Collapse -
(NT) Thanks a lot!
Jan 5, 2012 11:08PM PST
- Collapse -
(NT) Noted, and thanks.
Jan 6, 2012 4:25AM PST
- Collapse -
Answer
A layman's response
Jan 5, 2012 8:38PM PST

I can only give a layman's response because I lack the technical knowledge.

In fact it's difficult to equate single core speed with multi-core speed because of the way in which multi-core works.

Take my setup. Quad-core rated at 3.20 Ghz but each core is rated at 965 Hz. First glance multiplies 965 x 4 = 3860.

But it doesn't work like that. Each processor works independently of the other reading data and executing program instructions, (from Wikipedia), and so they can all work at the same time. So while individual processors are rated at slower speeds, the effect is that they can each work on different tasks, and so the user, (you and me), see the system working faster despite the apparent slowness of each processor.

Here's that Wiki page, if it helps; http://en.wikipedia.org/wiki/Multi-core_processor

Expect others to provide more detailed information.

Mark

- Collapse -
Thanks a lot!
Jan 5, 2012 11:09PM PST

<span id="INSERTION_MARKER">Helped me to get an idea.

- Collapse -
Answer
Try task manager
Jan 7, 2012 2:21AM PST

There is a good way to test the speed of your CPU and compare it with another CPU. If you have another computer around that's got a single core CPU, then you just simply turn on both computers and go into the task manager of both computers (ctrl-alt-delete) at the same time. When both task managers appear on the both screens, you will then select the PERFORMANCE tabs. From there you will be able to observe how fast the CPUs of both computers are running by looking at the CPU usage and memory meters and checking the rate of speed in the SYSTEM box or by clicking the resource monitor.

- Collapse -
For that to be meaningful
Jan 7, 2012 4:37AM PST

For that to be meaningful, you'd need both computers to have identical configurations, software and hardware, save the one thing you are testing. The more you deviate from that, the less accurate the figures you get, to the point where if you just try and compare two random desktops as you suggest, you can't draw any meaningful conclusions.

All you're seeing there is how much the CPU is being utilized as a percentage between 0 and 100% of it's capacity.

- Collapse -
Agree with Jmmy
Jan 7, 2012 4:49AM PST

I don't see it being helpful at all.

I don't even suppose bench testing would do it, Not only would the OS have to be identical, but so would the hardware, apart from the processors.

Mark

- Collapse -
Answer
How fast
Jan 7, 2012 5:34AM PST
- Collapse -
Answer
even more complex
Jan 9, 2012 5:54AM PST

As Jimmy mentioned, even with four cores, there are common data paths (and other resources that are not duplicated, times four) so a four core processor isn't necessarily four times faster than a single core processor of the same clock speed.

There are other considerations as well. Consider the software. Sure, one core might be handling an application while another does some operating system task, but it doesn't always work this way. Some tasks are fundamentally serial, not parallel, and cannot be divided between two or more cores. Think of one resource-hogging application that cannot be divided into smaller tasks. Instead of the highway analogy, think of it as a train analogy. One train with a hundred cars all connected won't carry any more cargo even if there are three additional empty tracks next to it.