st: what is a good processor type and working memory size to work with a large dataset in Stata 11?

[Evelyn Ersanilli <evelyn.ersanilli@qeh.ox.ac.uk>]

Dear stata-listers,

At our institute we are looking to buy new laptops to use in two projects that will compile and analyse a large datasets.
We are trying to figure out  what type of processor and what size working memory would be most appropriate, i.e. give an acceptable processing speed  (and still fit with our budget ;))
It would be great if you could give us some advice.

We estimate that dataset 1 will contain about 2 million observations (mostly numeric variables). For most analyses only a selection of observations will be used. The dataset will mainly be analysed with time-series techniques
Dataset 2 will contain about 8.000 cases with 500-600 variables so roughly 4 million observations (again very few string variables). Here we would often take all cases (but only a selection of variables) to do some fairly basic OLS, logit and multilevel analyses.
What would be a minimum processor speed and memory size to analyse these two datasets without having to go get a cup of tea every time we hit  the enter key?

I'm not quite sure what determines the speed of Stata calculations (other than of course the complexity of the analyses); the processor, working memory or the interaction of the two?
Could anyone explain this a bit?

We have a Stata 11 SE license.  Would it still make sense to get a duo or quad core processor?

Many thanks

Kind regards,

Evelyn


Research Officer 
EUMAGINE project 
 
International Migration Institute (IMI)
James Martin 21st Century School 
Department of International Development 
Mansfield Road
Oxford OX1 3TB
University of Oxford 
Tel: +44 (0)1865271925


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/