i5 16gb vs i7 8gb - same price?

[automata]

Hi everybody.

I am about to buy a new laptop for college and i have one question. I have searched on various websites for the past days, and found plenty of articles and such, but as we know its difficult to compare processors over different generetions. I don't know whether to buy the macbook pro 13" i5 with the upgraded 2.6 processor and 16gb ram - or the i7 2.8 with standard 8gb ram. It is absolutely the same price, but what will give me most power for my needs?

Right now i have the last of the white macbooks with 2.4 dual core and 2gb ram. I have monitored my task manager, and im lagging in both processor power as well as ram (constantly filled up). The little "beach-ball-waiting-logo" comes up constantly, even though im just typing this message (with 8 tabs in firefox + itunes running).

I am going to run some physics simulations, and use it for mathematical work at college. Nothing extreme though. I am also going to produce some low-end simple music through logic pro x, all together with listening to music through itunes (i have a big library). Then im going to be programming a bit php,c++,xcode for fun, with some borderlands 2 gaming and maybe counter strike or something like that for fun once in a while.

My old white macbook is really slowing me down, and i dont know whether to choose the new haswell 13" 2.6 ghz i5 with 16gb ram, or the 2.8 ghz i7 with 8 gb ram - it is the SAME price with edu-pricing. I do not want to buy the 15" because it's simply too big a screen.

Thank you very very much!

 

[prsuser90]

Same here, but with an additional question. Is a basic MacbookPro retina 15" the better bang for the buck ? (in comparison to the MacBooks mentioned above)

 

[leman]

If its simulations you are doing, then the i7 will improve your performance (16GB of RAM will not). However, its a good question to which margin. Maybe around 5%, depending on what you are doing.

P.S. If you have coded the simulations yourself, I advice you to recompile them for Haswell with AVX2 instruction set enabled - it could give you a healthy boost.

 

[Meister]

If its simulations you are doing, then the i7 will improve your performance (16GB of RAM will not). However, its a good question to which margin. Maybe around 5%, depending on what you are doing.

P.S. If you have coded the simulations yourself, I advice you to recompile them for Haswell with AVX2 instruction set enabled - it could give you a healthy boost.

This is good advice. For your purposes the i7 seems better.
More ram will do you no good.

 

[maflynn]

As already mentioned, CPU intensive stuff will benefit with the better CPU, then the 16GB of memory. In fact most computers don't need 16GB of ram right now, its more of future proofing since you cannot upgrade later. 8GB will be more then sufficient however so I'd go with i7 8gb model if I had to choose based on your stated needs.

 

[automata]

I will be running computations and simulations of various problems related to math and physics. It will be coded in matlab, wolfram mathematica, etc.

So you are saying that the i7 processor only will increase the power by 5% margin or so? Because i saw the benchmarks that primatelabs put out, and the increase was a bit more evident.

Can anybody tell me why 16gb will do me no good - or in which situation 16gb would be preferable? It's just because right now snow leopard is taking up my 2gb ram instantly - no matter what im doing. I am aware that a lot of things has changed since snow leopard, but it is really mind-blowing that 16gb wouldn't help more than 0.4 ghz going from i5 to i7

But I guess im in for the i7

 

[leman]

So you are saying that the i7 processor only will increase the power by 5% margin or so? Because i saw the benchmarks that primatelabs put out, and the increase was a bit more evident.

The primatelabs micro benchmark (multiple threads) is around 7% faster on the i7 compared to the i5. Expect less difference in real life usage.

Can anybody tell me why 16gb will do me no good - or in which situation 16gb would be preferable?

Well, if your datasets are exceeding several gigabytes in size (which I doubt), then 16Gb would be a good idea. I routinely worked with data mining on databases of around 1-2GB in size on my old 4GB MBP - without any noteworthy performance issues. My rMBP has 16GB RAM and its clearly too much. If I were to make a decision again I would have rather saved the money. I can tell because I have 8GB RAM on my iMac, and it does not show any RAM deficits on same workflows.

P.S. 8GB is 4 times of 2GB Increase of RAM size is a subject of diminishing returns, after a certain threshold your software can't really take advantage of it. Of course, it all depends on what you do - users working with a number of RAM-heavy applications (e.g. virtual machines) simultaneously will certainly see a difference.