Re: [boost] Accelerating algorithms with SIMD - Segmented iterators and alternatives

Subject: Re: [boost] Accelerating algorithms with SIMD - Segmented iterators and alternatives
From: Mathias Gaunard (mathias.gaunard_at_[hidden])
Date: 2010-10-16 14:01:12

On 15/10/2010 04:48, David Abrahams wrote:
 > At Thu, 14 Oct 2010 20:00:15 +0100,
 > Mathias Gaunard wrote:
 >>
 >> On 13/10/10 18:50, David Abrahams wrote:
 >>
 >>> I think you mean
 >>>
 >>>     1 2 | 3 4 5 6 | 7 8 9 10 | 11
 >>
 >> What is that nonsense supposed to be?
 >
 > I fear this conversation is becoming uncivil.

Sorry, I just couldn't make sense of what this is supposed to be, and 
used bad phrasing.

What would the value type of that range be?

 > That's what "if" statements are for.

We want to avoid these.

 >>> I think they do.  However, you might find that you need some dynamic
 >>> dispatching to get it to work.  In other words,
 >>>
 >>>     if (operating on complete segment)
 >>>        SIMD algorithm
 >>>     else
 >>>        elementwise algorithm
 >>
 >> Therefore they do not allow to avoid the inefficiency, and they are no
 >> different from standard iterators in that respect.
 >
 > Yes, they are different, because you don't need to do that test at the
 > lowest level, for each element.  That test would only be done at the
 > segment level.

How so?
The way I see if, I would need that if at every increment.

 > Unless you process the uneven end bits with non-SIMD code.

I would quite like a provide a way to do this.
A range, however, must be homogeneous, and have all its elements be the 
same type, so it's either a scalar or a vector, can't be both.