From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-163994-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 11966 invoked by alias); 1 Sep 2010 08:39:00 -0000
Received: (qmail 11951 invoked by uid 22791); 1 Sep 2010 08:38:58 -0000
X-SWARE-Spam-Status: No, hits=-1.9 required=5.0	tests=AWL,BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,RCVD_IN_DNSWL_NONE
X-Spam-Check-By: sourceware.org
Received: from mail-iw0-f175.google.com (HELO mail-iw0-f175.google.com) (209.85.214.175)    by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Wed, 01 Sep 2010 08:38:53 +0000
Received: by iwn2 with SMTP id 2so6797924iwn.20        for <gcc@gcc.gnu.org>; Wed, 01 Sep 2010 01:38:51 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.231.150.7 with SMTP id w7mr8784102ibv.14.1283330331608; Wed, 01 Sep 2010 01:38:51 -0700 (PDT)
Received: by 10.231.13.77 with HTTP; Wed, 1 Sep 2010 01:38:51 -0700 (PDT)
In-Reply-To: <201009010034.18924.paul@codesourcery.com>
References: <AANLkTimSiyGbOTCoSNb5q80YKtvjg46yrVxwbymi2qt=@mail.gmail.com>	<AANLkTim3vfWL5BsfeGg+VV2DOjnFRxkas2LT3waUQwQj@mail.gmail.com>	<AANLkTim=OiepFQ+-vqw1kJ_7zr+s9Sau1fBhoazABEWy@mail.gmail.com>	<201009010034.18924.paul@codesourcery.com>
Date: Wed, 01 Sep 2010 08:39:00 -0000
Message-ID: <AANLkTinUUndLEh5exJs+0_b6ueSBoJLWwUh3zMssYVAd@mail.gmail.com>
Subject: Re: Clustering switch cases
From: Richard Guenther <richard.guenther@gmail.com>
To: Paul Brook <paul@codesourcery.com>
Cc: gcc@gcc.gnu.org, "Paulo J. Matos" <pocmatos@gmail.com>, Ian Lance Taylor <iant@google.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
X-IsSubscribed: yes
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc/>
List-Post: <mailto:gcc@gcc.gnu.org>
List-Help: <http://gcc.gnu.org/ml/>
Sender: gcc-owner@gcc.gnu.org
X-SW-Source: 2010-09/txt/msg00002.txt.bz2

On Wed, Sep 1, 2010 at 1:34 AM, Paul Brook <paul@codesourcery.com> wrote:
>> > In fact we might want to move switch optimization up to the tree level
>> > (just because it's way easier to deal with there). =A0Thus, lower swit=
ch
>> > to a mixture of binary tree & jump-tables (possibly using perfect
>> > hashing).
>>
>> Doing the optimisation at the tree-level was exactly my initial idea.
>> By splitting the switches at the tree-level, before expand_case, would
>> then allow for expand_case to transform it either to a jump table or
>> binary tree depending on the situation.
>
> I'd kinda hope that doing the optimization at the tree level means expand=
_case
> doesn't have to handle both types. =A0The tree code converts sparse case =
ranges
> to explicit conditionals (or a switch on a compact perfect hash), so anyt=
hing
> left to RTL expansion must be a jump table.

Yes, that was my idea.

Richard.

> Paul
>