MBP Retina 8GB vs 16GB for data analysis - R, Python large files?

Question

Level 1

0 points

MBP Retina 8GB vs 16GB for data analysis - R, Python large files?

I increasingly find myself working with graph data (Neo4J, Gephi) and document databases (Mongo, CouchDB) as well as using R for clustering and analysis and Python to munge fairly large files of 1GB or larger.

8GB would probably serve me well and coming from a machine with 4GB will feel like plenty, but I suspect that I am the kind of use who would get his money worth out of 16GB.

In addition to this I will be running a Windows VM for Office Excel/VBA work and to stats packages for university very frequently.

Is getting the upgraded model worth it? Or will I not notice any difference.

Thanks,

David

Posted on Apr 3, 2013 8:16 AM

Reply

Answer 1

CT

Level 7

24,239 points

Apr 3, 2013 8:32 AM in response to semiChaotic

Yes, 16GB is worth it for intensive computational work such as you describe.

Reply

Answer 2

semiChaotic Author

Level 1

0 points

Jun 7, 2013 3:46 AM in response to CT

Good advice. I got the 16GB memory and when doing machine learning or social network analysis I see memory usage go beyond 8GB. I don't need it often but when I do it makes me more productive.

Reply

Answer 3

LowLuster

Level 6

12,105 points

Jun 7, 2013 3:56 AM in response to semiChaotic

I wouldn't be so sure the extra ram will help. The Program you are using might depend more on the cpu speed then the amount of ram. More ram only helps the system run at its top speed if you actually use up the currently installed amount of ram. So if any program does not use that extra ram, it does not need to place data into the extra ram, the system will not run any faster with more ram installed.

More is not always better. That's like saying If I had a billion dollars can I do more with 2 billion dollars? Yes and no, if you actually can use that extra billion.

Reply

Answer 4

Network 23

Level 6

12,351 points

Jun 7, 2013 7:25 AM in response to semiChaotic

I would say you will want the 16GB for a few reasons. You are working with memory-intensive datasets, and also you are running a virtual machine. VMs tend to use up a lot of RAM since it means you are running two complete OSs at the same time, on top of any other apps you might be running.

The third reason is that the Retina models cannot be upgraded after purchase. You would not want to have an 8GB forever if you discovered that your future work needed a bit more memory. You've already demonstrated doing some tasks that can be RAM-intensive, just max out the RAM and never worry about it again.

If you still have doubts, learn how to analyze RAM usage and particularly swap file behavior in Activity Monitor. If your Mac is already paging to swap files frequently, you already need the RAM. If you do work for several days and swap paging activity is low, you don't need the RAM.

Reply