This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Cost of inlining when the function has call to a target-specific builtin
- From: "Richard Guenther" <richard dot guenther at gmail dot com>
- To: "Pranav Bhandarkar" <pranav dot bhandarkar at gmail dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Sun, 16 Sep 2007 20:59:36 +0200
- Subject: Re: [PATCH] Cost of inlining when the function has call to a target-specific builtin
- References: <649555d50709160703x593bf9f3o5635ecb3fe8d38b9@mail.gmail.com>
On 9/16/07, Pranav Bhandarkar <pranav.bhandarkar@gmail.com> wrote:
> Hi,
> While inlining, the cost of a call to a target-specific builtin is
> assumed to be the same as that of a call to any other function by
> tree-inline.c. This patch allows the backend to examine the builtin
> and suggest an appropriate value.
This doesn't fit with the modularized cost metrics we have. That is,
estimate_num_insns_1 is not only used for inlining.
Richard.
> 2007-09-16 Pranav Bhandarkar <pranav.bhandarkar@celunite.com>
> Ramana Radhakrishnan <ramana.radhakrishnan@celunite.com>
>
> * target.h (inline_builtin_call): New hook.
> * expr.h (default_inline_builtin_call): New.
> * builtins.c (params.h): Include.
> (default_inline_builtin_call): New. Define.
> * target-def.h (TARGET_INLINE_BUILTIN_CALL): New
> (TARGET_INITIALIZER): Add TARGET_INLINE_BUILTIN_CALL.
> * tree-inline.c (estimate_num_insns_1): For CALL_EXPR check for
> target-specific builtin. If so get the cost for inlining from the backend.
>
>
> Is this ok ?
>
> cheers!
> Pranav
>
>