On 29-Jul-12 14:28, Era Scarecrow wrote:
On Sunday, 29 July 2012 at 09:30:06 UTC, Dmitry Olshansky wrote:
I have simpler suggestion. Maybe doing it as 2 containers:
BitSet is a plain value type with small array optimization (what you
called BitArray in last 2 posts) and no slicing (only as opSlice()
I.e. as a whole). Best used for small things under < 1000 bits (as
even dup will be cheap enough).
BitArray is class or struct that has reference smenatics and drops
small array optimization in favor of having simple & fast slicing Best
used for big arrays with no less then 100-200 bits.
As long as you only use one or the other then it sounds great; I'm
sure a conversion/constructor would be available to convert from one to
another. That does seem like the best solution, but also feels like
we're building it twice (Little code re-use).
Yup.
Might be a third option: A template struct that represents a BitArray
type (with 90% of the code); you then insert a structure with key
functions that would handle the value/reference & duplication part.
Sounds a lot like mixin template. I'd try to invent some small set of
primitive operations and auto-magically construct the others from it.
I'll meditate on it for a few days and see if anything pops up.
And just to keep your mind busy :) how about adding SIMD to speed it all up:
http://dlang.org/simd.html
It's kind of new & cool thing but not very well documented.
--
Dmitry Olshansky
