Re: Population count in SSE2, again

Maarten Kronenburg wrote:
> "Terence"  wrote in message
>> "James Van Buskirk"  wrote in message
>>  (code)
>>
>> Someone tell me why 21 instructions and a branch, per word, is somehow
>> better/faster than
>>
> 
> First I also didn't believe it, but somehow array elements are combined
into
> one element, and only that one element is counted, it's from:

There's no "somehow" at all about it:

This is a classic bitslice implementation of a FULL_ADD, i.e. 3 inputs 
and 2 outputs.

Terje

-- 
- <Terje.Mathisen@[EMAIL PROTECTED]
>
"almost all programming can be viewed as an exercise in caching"