Re: [PATCH] Replace completions with semaphores

[Matthew Wilcox <matthew@wil.cx>]

On Mon, Apr 14, 2008 at 05:41:29PM +0200, Ingo Molnar wrote:
> * Matthew Wilcox <matthew@wil.cx> wrote:
> > On Sun, Apr 13, 2008 at 09:05:13AM +0200, Ingo Molnar wrote:
> > > there's also another aspect: completions are faster a bit 
> > > in theory, because they know that they will schedule most of the time - 
> > > while semaphores assume that they will _not_ schedule. (And that's 
> > > exactly because the intent of the developer when using a completion is 
> > > crystal clear.)
> > 
> > In practice though, the current implementation is slower. [...]
> 
> any URL to benchmarks?

No -- just reading the code.

> > [...] Of course, that's fixable, and I strongly suspect that the 
> > current users of completions simply don't care about speed -- the 
> > normal use of completions is in startup and shutdown paths where a 
> > millisecond extra isn't going to be noticable.
> 
> completions and semaphores act in the sub-microsecond range, not in the 
> milliseconds range.

I've clearly misled both you and Jens, for which I apologise.  My point
was "even if they were a millisecond slower, nobody would notice",
rather than "I've measured it and they're a millisecond slower" or even
"from eyeballing it, I estimate they're about a millisecond slower".

-- 
Intel are signing my paycheques ... these opinions are still mine
"Bill, look, we understand that you're interested in selling us this
operating system, but compare it to ours.  We can't possibly take such
a retrograde step."