On Tuesday, September 25, 2012 4:22:03 PM UTC-4, William wrote: > > On Tue, Sep 25, 2012 at 1:14 PM, William Stein > <wst...@gmail.com<javascript:>> > wrote: > > On Tue, Sep 25, 2012 at 1:10 PM, rickhg12hs > > <rickh...@gmail.com<javascript:>> > wrote: > >> Is there a way to use R from Sage to speed things up? Is there > something in > >> Sage that could be improved to approach R's speed? > >> > >> E.g., here's a toy demonstration of how slow R's summary is from Sage. > >> > >> sage: version() > >> 'Sage Version 5.3, Release Date: 2012-09-08' > >> sage: %timeit r.summary(range(1000)) > >> 5 loops, best of 3: 17.5 s per loop > > > > I don't understand your questions above at all, but want to point out > > that in your example the time is completely dominated by converting > > the Python object "range(1000)" to R. Consider: > > > > sage: %timeit r.summary(range(1000)) > > 5 loops, best of 3: 2.2 s per loop > > sage: s = r(range(1000)) > > sage: %timeit r.summary(s) > > 125 loops, best of 3: 5.26 ms per loop > > Here is doing something similar (but not identical) in Sage (pure Python): > > sage: v = range(1000) > sage: %timeit [min(v), median(v), mean(v), max(v)] > 625 loops, best of 3: 121 µs per loop > > Here's something related: > > sage: v = stats.TimeSeries(range(1000)) > sage: %timeit [v.min(), v.mean(), v.max()] > 625 loops, best of 3: 3.67 µs per loop > > I'm sure R's summary is on the order of microseconds as well, and that > 5.26ms is almost all overhead. One can use the C interface to R to > get much better performance: > > sage: import rpy2.robjects as robjects > sage: v = robjects.r(range(1000)) > sage: summary = robjects.r['summary'] > sage: %timeit summary(v) > 625 loops, best of 3: 898 µs per loop > sage: print(summary(v)) > Min. 1st Qu. Median Mean 3rd Qu. Max. > 999 999 999 999 999 999 > > > So with the C interface it is 898 microseconds, which to me still > seems like a lot, given that I feel like I could write something in > Cython that does the same thing in about 10 microseconds. > > -- William >
You're right. Converting the Python object to R does dominate. I'm intrigued by your use of rpy2.robjects and why the summary is surprising. Regards. -- You received this message because you are subscribed to the Google Groups "sage-support" group. To post to this group, send email to sage-support@googlegroups.com. To unsubscribe from this group, send email to sage-support+unsubscr...@googlegroups.com. Visit this group at http://groups.google.com/group/sage-support?hl=en.
