From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-163471-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 23734 invoked by alias); 28 Jul 2010 18:17:04 -0000
Received: (qmail 23724 invoked by uid 22791); 28 Jul 2010 18:17:03 -0000
X-SWARE-Spam-Status: No, hits=-2.0 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-ww0-f51.google.com (HELO mail-ww0-f51.google.com) (74.125.82.51)    by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Wed, 28 Jul 2010 18:16:57 +0000
Received: by wwc33 with SMTP id 33so102843wwc.8        for <gcc@gcc.gnu.org>; Wed, 28 Jul 2010 11:16:54 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.216.169.80 with SMTP id m58mr10970922wel.79.1280341014532; 	Wed, 28 Jul 2010 11:16:54 -0700 (PDT)
Received: by 10.216.133.84 with HTTP; Wed, 28 Jul 2010 11:16:54 -0700 (PDT)
In-Reply-To: <824586E7387D1443B5DD024DAC75B91C139049@SE000319.ztb.icb.commerzbank.com>
References: <824586E7387D1443B5DD024DAC75B91C139048@SE000319.ztb.icb.commerzbank.com>	<AANLkTinSN+dOi-oi1ZWBa+4P55+Wvn=DtPasMJ_usOUb@mail.gmail.com>	<824586E7387D1443B5DD024DAC75B91C139049@SE000319.ztb.icb.commerzbank.com>
Date: Wed, 28 Jul 2010 18:17:00 -0000
Message-ID: <AANLkTi=5t20XgJ-m4Zbe3F2f7OUBaHOEfkxoHK27pF8R@mail.gmail.com>
Subject: Re: optimization question: mpl
From: Piotr Rak <piotr.rak@gmail.com>
To: "Hite, Christopher" <Christopher.Hite@partner.commerzbank.com>
Cc: Richard Guenther <richard.guenther@gmail.com>, gcc@gcc.gnu.org
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: 2010-07/txt/msg00406.txt.bz2

Hi,

2010/7/28 Hite, Christopher <Christopher.Hite@partner.commerzbank.com>:
>> Generally without knowing the compiler version you are using
>> it is hard to tell.
> I'll use whatever's best. =A0Right now I'm still on 4.4.3. =A0I'll probab=
ly
> upgrade soon to 4.5.
>
>> The same is true without a complete compilable testcase.
> I didn't want to make a test case that depends on boost::mpl. How's
> this:
>
> struct DecodeContext;
>
> struct M1{
> =A0 =A0 =A0 =A0static const int id=3D1;
> =A0 =A0 =A0 =A0static void decode(DecodeContext& ){}
> };
>
> struct M2{
> =A0 =A0 =A0 =A0static const int id=3D2;
> =A0 =A0 =A0 =A0static void decode(DecodeContext& ){}
> };
>
> struct M3{
> =A0 =A0 =A0 =A0static const int id=3D3;
> =A0 =A0 =A0 =A0static void decode(DecodeContext& ){}
> };
>
>
> template <int> struct ListMember;
>
> template <> struct ListMember<1>{ typedef M1 type; };
> template <> struct ListMember<2>{ typedef M2 type; };
> template <> struct ListMember<3>{ typedef M3 type; };
>
> template<int i>
> void foo(int id, DecodeContext& dc) {
> =A0 =A0 =A0 =A0typedef typename ListMember<i>::type M;
> =A0 =A0 =A0 =A0if(M::id=3D=3Did)
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0M::decode(dc);
> =A0 =A0 =A0 =A0else
> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0foo<i-1>(id,dc);
> }
>
> template<>
> void foo<0>(int id, DecodeContext& dc) {}
>
>
>
> int main(){
> =A0 =A0 =A0 =A0DecodeContext& dc=3D *(DecodeContext*)0;// junk for now
> =A0 =A0 =A0 =A0int id=3D2; //sometime runtime dependent
> =A0 =A0 =A0 =A0foo<3>(id,dc);
> =A0 =A0 =A0 =A0return 0;
> }
>
>
>> You can use the flatten attribute to tell the compiler to inline all
>> calls in a given function, like
>>
>> void __attribute__((flatten)) foo(void)
>> {
>> ...
>> decode1();
>> ...
>> }
>
> That would cause decode1() to be inlined, which might not be what you
> want.
>
> Hmm maybe I could rewrite things so the switch case returns a function
> pointer. =A0I'm guessing that would make things slower though.
>

Or you could just initialize static array of pointers, like that
(please note, that I never compiled code below):

typedef void (*pfn) (DeclContext&);

tempate <int I>
struct InitDispatchHelper
{
  static void init(std::tr1::array<pfn,N>& a)
  {
    a[I-1] =3D ListMemeber<I-1>::foo;
    InitPointersHelper<I-1>::init(a);
  }
};

// End recursion
template<>
struct InitDispatchHelper
{
  static void init(std::tr1::array<pfn,N>& a)
  {    /*does nothing */  }
};

// Dispatch table;
template <int N>
struct DispatchFoo : std::tr1::array<pfn, N>
{
  DispatchFoo() : std::tr1::array<pfn, N>()
  {
     InitDispatchHelper<N>::init(*this);
  }
};

then your call would look like:

static const int X =3D /* your maximal I in ListMembers meta-class */

void call_foo(int i, DeclContext& c)
{
  static const DispatchFoo<X> foo_dispatch;
  foo_dispatch[i] (c);
}

Bonus points for benchmarking both solutions and making it policy
class choosing between those two solutions depending on X.

Good luck,

Piotr