This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: [PATCH] Fix fortran creating a huge number of io transfer structs
- From: "François-Xavier Coudert" <fxcoudert at gmail dot com>
- To: "Richard Guenther" <rguenther at suse dot de>
- Cc: gcc-patches at gcc dot gnu dot org, fortran at gcc dot gnu dot org
- Date: Wed, 22 Feb 2006 18:15:17 +0100
- Subject: Re: [PATCH] Fix fortran creating a huge number of io transfer structs
- Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=beta; d=gmail.com; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=U5T5WxZR7rPBpkgFodjQYiK/4Lg5qFICyr4huFw8QZfs8syCvxrhnvlmA7n+AOBoJ7wdIj36ee/xkPzh1K2l/4nJf0F7kcnExTb17ximOHKxgwf2FrHZoEeusA/0ExakMbGXYB61xVSOPYx6g4s+WZjrBv1ichf0uTSOWK5Z7ZU=
- References: <Pine.LNX.4.63.0602221629100.7478@t148.fhfr.qr>
> PROGRAM test
> WRITE(*,*) 'hello'
> WRITE(*,*) 'world'
> END
>
> gfortran creates two struct __st_parameter_dt, one for each I/O operation.
> This causes a _huge_ regression in compile time if we ever start to create
> SFTs for those structs
Hum, could we have a pointer to what SFT is? (a google "sft gcc" lead
me to info page
http://gcc.gnu.org/onlinedocs/gccint/Alias-analysis.html, but that's
still not completely obvious to my simple mind...)
> but there's no reason why gfortran cannot share one struct
> __st_parameter_dt per function or at least one per sequence of I/O
> operations.
I wonder how that does interact with the fact asynchronous I/O
(although it's not implemented yet, we still have to think about it).
FX