On Fri, May 25, 2012 at 8:00 PM, Jorge <[email protected]> wrote:

> On May 26, 2012, at 1:35 AM, Nathan Rajlich wrote:
>
> > Jorge, what confuses me is that you're pointing out the same bug that
> you discovered in another thread from a while back (when you were claiming
> that your threads beat vanilla node). mraleph quickly discovered and fixed
> the bug, IIRC.
> >
> > So still, I don't know what point you're trying to make...
>
> Nathan, in the thread "When threads_a_gogo (*) beats node 2 to 1" a
> thread_a_gogo was 2x times faster that node's main thread, but now, here,
> what we're seeing is that Apple's jsc is running that über-simple loop ~
> 50x faster than node! Even after applying Egorov's patch, node would still
> be 25x slower, and that's worrying (IMO), isn't it?
>

No because that code wasn't the same code AT ALL.

Clearly there's a missing optimisation step in the current node - not a
de-facto 2x speedup, but a way of running code that would miss the
optimiser completely, possibly resulting in pathological cases in a 25x
slowdown for all we know. Clearly that has been fixed in V8, given other
people's benchmarks of "./shell". Have you actually run ./shell from
bleeding_edge to test this against? Other people appear to have and
demonstrate that your pathological case has been fixed.

Matt.

-- 
Job Board: http://jobs.nodejs.org/
Posting guidelines: 
https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines
You received this message because you are subscribed to the Google
Groups "nodejs" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Reply via email to