On Thu, Sep 16, 2010 at 12:31 PM, Frank Mittelbach
<[log in to unmask]> wrote:
> Chris Rowley writes:
>  > A vaguer question: Is the level of the relatively efficiency of the types
>  > of register of any practical consequence in the 21st century?
> I was wondering about that too, guess the only way to find out is to make some
> experiments

I carried out some experiments with this a few years ago and found the
count registers from the extended pool to be only slightly slower (5%
or so) than the ones from the normal pool.