clock speed and cores

10 views (last 30 days)
William
William on 10 Mar 2011
Answered: John D'Errico on 16 Sep 2019
Hi all, some of the answers from my last question helped quite a bit and I have narrowed down my laptop selection. I realize that these kind of questions may be difficult to answer because different uses of MATLAB will benefit from different hardware setups etc.. but any insight, even general, would be appreciated.
I am choosing between two mac books
13" 2.7ghz i7 Duo w/Intel HD Graphics 3000
or
15" 2.0ghz Quad w/AMD Radeon HD 6490M with 256MB GDDR5
Currently I use MATLAB for mostly integration of ODEs and PDEs. Often in large dimensions.
I do not currently use the parallel computing toolbox, but would like to learn how to use it effectively if it applies to my work..in other words - it may be in my future.
I am wondering if the 2.7ghz may run faster, in general, than the 2.0ghz, if the clock speed will have more impact than the cores.
I know that I am oversimplifying the question and that multiple factors will affect speed...and maybe it will come down to just having a bigger screen to work with.
I am leaning towards the 15"
Any insight would be very appreciated, thank you!

Answers (1)

John D'Errico
John D'Errico on 16 Sep 2019
In general, this is impossible to answer. For ultimate speed, you want many cores, a fast clock speed, fast disk access, and a good system fan. Other factors too, such as cache memory, graphics card, etc.
Which is more important? Again, impossible to know, because it will fully depend on what you will be doing, EXACTLY which problems you will be solving. In fact, there are often subtly different schemes and algorithms you might use to optimize your code, depending on your hardware. So your choice of hardware might easily influence your code.
Essentially, get the best computer you can afford. I would point out that IF you were truly looking for optimal performance, a laptop system might not be your best choice, as systems will often see performance degradation under heavy use if they heat up. The CPU necessarily slows down to avoid heating problems. This will happen when you are running flat out all of those cores the manufacturer pushes on you, convincing you that you are getting true power. The fan then often does not suffice, your CPU gets hot, and your clock speed automatically throttles down. The point is, the fastes computer you can get for flat out performance will have a superb cooling system, and that will not fit well in a small laptop case. So if you really want performance, then the biggest source of impact on ultimate performance may well be the fact that you chose a laptop in the first place.

Categories

Find more on Parallel Computing Fundamentals in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!