On Feb 28, 2005, at 1:36 PM, Alan Hudson wrote:
> Paul Sandoz wrote:
>
>>
>>> Have you verified this recently.
>>
>>
>> I did some tests a while back and final made a measured difference,
>> not a large one but nothing to be sniffed at :-)
>>
> good enough then. I just didn't want to be going on old data. Ie it
> used to really make a difference... guess it still does. I might
> timetrial this myself sometime again as I had thought the issue went
> away.
Last time I checked, it was still useful to use final for
high-performance code. The runtime can use that information as an
inlining hint too.
-- Santiago