Most likely cause is that you're using non-constant global variables in your code, but as Milan said, it's hard to tell without seeing the code.
On Sun, Jan 25, 2015 at 11:49 AM, Milan Bouchet-Valat <[email protected]> wrote: > Le dimanche 25 janvier 2015 à 08:45 -0800, Martin Kapfhammer a écrit : > > I generated a data set > > 10 000 000 rows á 2 elements > > resulting in 476 MB > > > > > > and cluster it using kmeans algorithms with k = 3. > > > > > > > > > > Julia over 5 minutes > > R 20 seconds > > Scipy 16 seconds > > > > > > My machine has 12 cores and 64 MB memory. > > > > > > Why are they so fast? > Hard to tell. Could you share the code you used (for all the languages > if possible)? > > > Are R and scipy clustering in parallel? > You should be able to check that by looking at CPU use when running > these commands. > > > Regards > >
