1. Unoptimized slicing?
- Posted by akusaya at gmx.net
Jan 01, 2004
a = rand(repeat(200, 200))
I found that the operation
c = a[1..length(a)]
or
c = a[1..200]
is much slower than
c = a
Why isn't it optimized? (check the length of the slice, if it's equal
to the original sequence the pointer should be copied not the
contents)
2. Re: Unoptimized slicing?
aku saya wrote:
> a = rand(repeat(200, 200))
>
> I found that the operation
>
> c = a[1..length(a)]
> or
> c = a[1..200]
>
> is much slower than
>
> c = a
>
> Why isn't it optimized? (check the length of the slice, if it's equal
> to the original sequence the pointer should be copied not the
> contents)
Thanks. I'll think about it.
With a special case optimization like this,
you waste a bit of time checking for the condition
hundreds, or maybe thousands, of times for each
time it actually occurs.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
