Will Coleda wrote:
On Tue, Aug 26, 2008 at 10:53 AM, luben <luben@unixsol.org> wrote:
Is this on purpose? The implicit return is 4-5 times faster than explicit
return.
Best regards
luben
CC'ing perl6-compiler (where rakudo-particular items should go), and
wondering if you can attach the generated PIR for those of us who are
interested but unwilling at the moment to sync up and try this out.
Thanks. ^_^
In order to measure the performance difference I am using a dumb
Fibonacci numbers calculation:
sub fib ( $n ){
($n < 2)?? $n !! fib($n-1)+fib($n-2);
}
say(fib(22));
some measurmens:
NQP explicit return : 2.686s
NQP implicit return : 0.555s
perl6 explicit return: 18.518s
perl6 implicit return: 4.768s
luben
From: Erik Johansen <erik@adapt.dk>
X-Virus-Check-By: mailtest2.pair.com
X-Spam-Check-By: mailtest2.pair.com
X-Spam-Status: No, hits=-9.5 required=4.0
tests=BAYES_00,DATE_IN_PAST_06_12,RCVD_IN_DNSWL_HI autolearn=ham
version=3.002004
X-Spam-Flag: NO
X-Spam-Level:
X-Spam-Filtered: 052f96e22adcb440ba93d277d4a1272f
Mailing-List: contact perl6-all-help@perl.org; run by ezmlm
Precedence: bulk
List-Post: <mailto:perl6-all@perl.org>
List-Help: <mailto:perl6-all-help@perl.org>
List-Unsubscribe: <mailto:perl6-all-unsubscribe@perl.org>
List-Subscribe: <mailto:perl6-all-subscribe@perl.org>
List-Id: <perl6-all.perl.org>
Delivered-To: mailing list perl6-all@perl.org
Delivered-To: perl6-all-poster@perl.org
X-Mailing-List: contact perl6-internals-help@perl.org; run by ezmlm
X-Mailing-List-Name: perl6-internals
List-Id: <perl6-internals.perl.org>
Delivered-To: mailing list perl6-internals@perl.org
Delivered-To: moderator for perl6-internals@perl.org
Delivered-To: perl6-internals@perl.org
Delivered-To: parrotbug-followup@parrotcode.org
Organization: Adapt A/S
To: parrotbug-followup@parrotcode.org
Subject: Re: [perl #57920] [PATCH] Suggestion for Parrot Configure test of AIO
Date: Wed, 27 Aug 2008 08:19:48 +0200
User-Agent: KMail/1.9.9
References: <RT-Ticket-57920@perl.org>
<rt-3.6.HEAD-29759-1219756519-1667.57920-74-0@perl.org>
<rt-3.6.HEAD-29762-1219770302-1565.57920-74-0@perl.org>
In-Reply-To: <rt-3.6.HEAD-29762-1219770302-1565.57920-74-0@perl.org>
MIME-Version: 1.0
Content-Type: text/plain;
charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Message-Id: <200808270819.48825.erik@adapt.dk>
X-Old-Spam-Check-By: la.mx.develooper.com
X-Old-Spam-Status: No, hits=1.5 required=8.0
tests=BAYES_00,RDNS_NONE,URIBL_BLACK
On my OpenSuse, it looks like this:
auto::aio - Does your platform support AIO...wrong signal
=2E.................done.
So the fact that it exits the loop early, just makes it continue to the "wr=
ong=20
signal" indication, and that marks it as not supported. After all it did no=
t=20
meet the test to return the expected signal.
/Erik Johansen
On Tuesday 26 August 2008, James Keenan via RT wrote:
On Tue Aug 26 06:15:19 2008, erik@adapt.dk wrote:
> On Monday 25 August 2008, James Keenan via RT wrote:
> > 1. After you run 'perl Configure.pl' with this patch on your box,
>
> what
>
> > do you get for this:
> >
> > grep -ni AIO lib/Parrot/Config/Generated.pm
>
> AIO is not in the file.
>
> (I am not sure that this is the right conclusion. After all, opensuse
> has /lib/libaio.so.1 & /lib/libaio.so.1.0.1 indicating that AIO is
> supported. It may be more likely that the AIO works fine, but the
> signal gets
> sent before the handler is set up. And that makes it sit waiting for a
> signal
> that does not arrive. So a rewrite of the test program might be the
> real
> solution.)
Yes, this is a bit puzzling. With your patch applied, calling 'perl
Configure.pl' gives me this output:
auto::aio - Does your platform support AIO......................yes.
... and doing the grep I described above gives:
grep -in aio lib/Parrot/Config/Generated.pm
92: 'HAS_AIO' =3D> 1,
110: 'aio' =3D> 'define',
... and this is on two boxes (one Ubuntu, one Debian) where I'm sure I
never did anything special to install AIO.
And so even with your patch applied, when you run Configure.pl you don't
detect the presence of AIO -- am I understanding you correctly?
The probe file in question has not been substantively altered in four
years, so if you feel it should be rewritten, please feel free to submit
a different patch.
HTH
kid51
=2D-=20
Med venlig hilsen
Erik Johansen, Innovation & udvikling
Adapt A/S, Langebrogade 6E, 3. sal, 1411 K=C3=B8benhavn K, Tlf. +45 3341 10=
50
|