bug-apl
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-apl] 100?100 hang-up


From: Juergen Sauermann
Subject: Re: [Bug-apl] 100?100 hang-up
Date: Wed, 14 Oct 2015 13:32:53 +0200
User-agent: Mozilla/5.0 (X11; Linux i686; rv:31.0) Gecko/20100101 Thunderbird/31.4.0

Hi,

thanks, fixed in SVN 683.

/// Jürgen


On 10/13/2015 12:53 PM, Kacper Gutowski wrote:
On Tue, Oct 13, 2015 at 12:17 PM, Elias Mårtenson <address@hidden> wrote:
It should be new set_size/8+1 on both the like with new and the subsequent
memset() one.
You mean (set_size+7)/8 :)

I was almost expecting this to be related to that this algorithm's
termination probability converges to 1 very slowly when A is close to B,
but 100 is small and 100?100 worked for me from time to time regardless
of ⎕RL so that's not it.  Good catch, Jay!


While this isn't related to the bug reported, this algorithm saves a lot
of memory when A << B, but otherwise it doesn't buy us anything because
A cells need to be allocated for the result anyway.  I think it might
be a good idea to switch to standard Fisher-Yates shuffle at some point
when A gets close to B (i.e. allocate B IntCells, shuffle in place and
return first A of them; this terminates deterministically in linear time).

-k




reply via email to

[Prev in Thread] Current Thread [Next in Thread]