Rob,

You're right about the timing. I had it backwards.

On Fri, Jun 2, 2017 at 10:45 AM, robert therriault <[email protected]>
wrote:

> Hi Brian,
>
> Your implementation of Henry's suggestion doesn't look faster to me.
>
>     2.84443e_5 % 3.79653e_6  NB. time for trimhb % time for trimalt
> 7.49218
>
> From a more recent post, Henry's suggestion was
>
> tess&(cp;._3"2)&.(2 0 1&|:)i.7 10 3
>
>     10000 timespacex 'tess&(cp;._3"2)&.(2 0 1&|:)i.7 10 3'
> 1.81361e_5 34304
>
>  trimalt=: {~ <@:(>:@:(3 * i.)@:<.@:(%&3)&.>)@:}:@:$
>    10000 timespacex 'trimalt i. 7 10 3'
> 2.1667e_6 7040
>
>
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to