From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-patches-return-131030-listarch-gcc-patches=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 7211 invoked by alias); 16 Oct 2004 18:35:00 -0000
Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc-patches/>
List-Post: <mailto:gcc-patches@gcc.gnu.org>
List-Help: <mailto:gcc-patches-help@gcc.gnu.org>
Sender: gcc-patches-owner@gcc.gnu.org
Received: (qmail 7187 invoked from network); 16 Oct 2004 18:34:59 -0000
Received: from unknown (HELO mx1.redhat.com) (66.187.233.31)
  by sourceware.org with SMTP; 16 Oct 2004 18:34:59 -0000
Received: from int-mx1.corp.redhat.com (int-mx1.corp.redhat.com [172.16.52.254])
	by mx1.redhat.com (8.12.11/8.12.10) with ESMTP id i9GIYopW002253;
	Sat, 16 Oct 2004 14:34:55 -0400
Received: from devserv.devel.redhat.com (devserv.devel.redhat.com [172.16.58.1])
	by int-mx1.corp.redhat.com (8.11.6/8.11.6) with ESMTP id i9GIYmr28167;
	Sat, 16 Oct 2004 14:34:48 -0400
Received: from devserv.devel.redhat.com (localhost.localdomain [127.0.0.1])
	by devserv.devel.redhat.com (8.12.11/8.12.10) with ESMTP id i9GIYPFC001673;
	Sat, 16 Oct 2004 14:34:25 -0400
Received: (from jakub@localhost)
	by devserv.devel.redhat.com (8.12.11/8.12.11/Submit) id i9GIYPbt001671;
	Sat, 16 Oct 2004 14:34:25 -0400
Date: Sat, 16 Oct 2004 18:37:00 -0000
From: Jakub Jelinek <jakub@redhat.com>
To: Richard Henderson <rth@redhat.com>, Matt Austern <austern@apple.com>,
        GCC Patches <gcc-patches@gcc.gnu.org>
Subject: Re: [Committed] Use special-purpose hash table to speed up walk_tree
Message-ID: <20041016183424.GC31909@devserv.devel.redhat.com>
Reply-To: Jakub Jelinek <jakub@redhat.com>
References: <F96640A6-1E36-11D9-9854-000A95AA5E5E@apple.com> <20041016094708.GZ31909@devserv.devel.redhat.com> <20041016103712.GA31909@devserv.devel.redhat.com> <20041016182952.GA31659@redhat.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20041016182952.GA31659@redhat.com>
User-Agent: Mutt/1.4.1i
X-SW-Source: 2004-10/txt/msg01344.txt.bz2

On Sat, Oct 16, 2004 at 11:29:52AM -0700, Richard Henderson wrote:
> On Sat, Oct 16, 2004 at 06:37:12AM -0400, Jakub Jelinek wrote:
> > +   choice for A is the integer part of (ULONG_MAX + 1.0) / phi, where phi
> > +   is the golden ratio.
> 
> Any reason not to use the computation actually described here?

You mean doing the computation in FPU?

> Otherwise, the ifdef one's fine if you use HOST_BITS_PER_LONG
> to choose one of the alternatives exactly:
> 
> #if HOST_BITS_PER_LONG == 32
>   const unsigned long A = 0x9e3779b9u;
> #elif HOST_BITS_PER_LONG == 64
>   const unsigned long A = 0x9e3779b97f4a7c16ul;
> #endif
> 
> Which will handily not compile if we ever find something else.

Yeah, that looks better.

	Jakub