The results listed on the page (i.e. that "."0@": is the fastest algorithm
for computing digits in 2^10000x) is absolutely correct--I checked it--but
it should be stated that 10^#.^:_1 is faster for large arrays of short
numbers.

Marshall

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of R.E. Boss
Sent: Tuesday, August 31, 2010 9:33 AM
To: 'Programming forum'
Subject: Re: [Jprogramming] Splitting an integer into its digits

> Van: [email protected] [mailto:programming- 
> [email protected]] Namens Marshall Lochbaum

(...)
> Since 10&#.^:_1 is indeed faster than "."0@":"0 (which I did not 
> expect, as the latter is listed under special code), this gives the 
> fastest execution yet.
(...)


   rank~ '10&#.^:_1 d';'"."0@": d'[d=:i.1e5
+-----------+----+-------+-----+------+
|expression |rank|tm*sz  |time |size  |
+-----------+----+-------+-----+------+
|10&#.^:_1 d| 0  |   1.00| 1.00|  1.00|
+-----------+----+-------+-----+------+
|"."0@": d  | 1  |2374.99|20.12|118.05|
+-----------+----+-------+-----+------+

This result indeed needs amending
http://www.jsoftware.com/help/release/digits10.htm 


R.E. Boss

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to