So, what is the answer to Alessandro's question, namely, whether any of the array-based solutions avoid redundant computations? We know the looping approach avoids redundancy, but what about the others?
I'm not enough of a J expert to answer that question, as it requires a more intimate knowledge of the internals and optimizations of the J primitives than I have. One of the interesting effects of array languages is that the interpreter can detect and optimize array operations "under the covers", without the programmer having to worry about those optimizations. The J sort operator /: is a good example of this type of phenomena. Skip On Sat, Dec 22, 2012 at 9:52 AM, Boyko Bantchev <boyk...@gmail.com> wrote: > On 21 December 2012 01:02, Skip Cave <s...@caveconsulting.com> wrote: > > The best way to answer his question is to program the problem using his > > explicit looping approach, and compare that approach with the approaches > > proposed by the various members of the forum. > > Comparing solutions with respect to, say, speed, may only tell us > which one to use (once we have them all!) if we want that speed. > But such a comparison does not answer Alessandro's question, namely, > whether array-based solutions, or a particular one of them, avoid > the redundant computations that he wrote of. > > Furthermore, comparing solutions with explicit loops to solutions > without them always tends to be unfair to the former. Huge differences > in speed do arise not only because J performs excelently on whole- > array operations, but also because it is very slow when it has to > execute code repeatedly. In this it is much, much slower than most > other languages. > > This is easily demonstrated by writing in another language a program > similar to the above discussed acx, such as the following program in > Ruby: > > n = 5000 > a,b = 1,0 > r = [[a,b]] > (n-1).times do > a,b = 5*a,5*b+3 > r << [a,b] > end > > Regarding performance, Ruby is well known for the lack of it. Still, > this program is faster by orders of magnitude than acx. It is also > faster than hr1 and hr2, although not much faster – only about twice > on my computer for n=5000. (The speed ratio increases with the growth > of n.) > > One more observation. How can one *predict* that the closed-formula > based solution of Bo and Linda will lose so dramatically to Henry's > solutions? > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > -- Skip Cave Cave Consulting LLC ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
