4.3 release plan

4.3 release plan

[Bernardo Innocenti]

Hello,

I know it's way early to ask, but has a release plan
already been set for the 4.3 release?  A gross estimation
of the expected release date would be sufficient.

I'm asking because I need to decide whether I need to do
the effort to backport a few things to 4.2 or I can rely
on the compiler being at least close to release when I'll
need it to be stable.

-- 
   // Bernardo Innocenti
 \X/  http://www.codewiz.org/

Re: 4.3 release plan

[Mike Stump]

On May 18, 2007, at 12:51 PM, Bernardo Innocenti wrote:
> has a release plan already been set for the 4.3 release?

Just take the dates between 4.1 and 4.2, and add to the 4.2, and  
presto, you have the 4.3 times...  Or, put another way, about 15  
months from now.[1]

1 - YMMV, YGWYPF, PPINGOFR

Re: 4.3 release plan

[Bernardo Innocenti]

Mike Stump wrote:
> On May 18, 2007, at 12:51 PM, Bernardo Innocenti wrote:
>> has a release plan already been set for the 4.3 release?
> 
> Just take the dates between 4.1 and 4.2, and add to the 4.2, and  
> presto, you have the 4.3 times...  Or, put another way, about 15  
> months from now.[1]

Come on, 4.3 doesn't look in such a bad shape!  And not long
ago there were people proposing to skip 4.2 and release 4.3
directly.


> 1 - YMMV, YGWYPF, PPINGOFR

I got the first two, but what does the last one mean?


BTW, the tentative timeline says that 4.3 stage 1 will end
4 months *ago*:
  http://gcc.gnu.org/develop.html#timeline

-- 
   // Bernardo Innocenti
 \X/  http://www.codewiz.org/

Re: 4.3 release plan

[Steven Bosscher]

On 5/18/07, Bernardo Innocenti <bernie@codewiz.org> wrote:
> Come on, 4.3 doesn't look in such a bad shape!

It will soon, when the stage1 projects are finally merged into the trunk.

> BTW, the tentative timeline says that 4.3 stage 1 will end
> 4 months *ago*:
>   http://gcc.gnu.org/develop.html#timeline

Being tentative, the timeline tends to be wrong ;-)

Gr.
Steven

Re: 4.3 release plan

[Mike Stump]

On May 18, 2007, at 2:45 PM, Bernardo Innocenti wrote:
> Come on, 4.3 doesn't look in such a bad shape!

I'll let history decide...

>> 1 - YMMV, YGWYPF, PPINGOFR
>
> I got the first two, but what does the last one mean?

Wow, you're impressive...  The second one I thought would be too  
cryptic for all.  You didn't try google did you?  :-)  Past  
performance is no guarantee of future returns.

> BTW, the tentative timeline says that 4.3 stage 1 will end 4 months  
> *ago*:
>   http://gcc.gnu.org/develop.html#timeline

Wait, this seems to contradict your point #1 above.

Re: 4.3 release plan

[Paolo Bonzini]

M
>>> 1 - YMMV, YGWYPF, PPINGOFR
>>
>> I got the first two, but what does the last one mean?
> 
> Wow, you're impressive...  The second one I thought would be too cryptic 
> for all.

I got the second one too.  Italians must be good at acronyms. :-)

>> BTW, the tentative timeline says that 4.3 stage 1 will end 4 months 
>> *ago*:
>>   http://gcc.gnu.org/develop.html#timeline
> 
> Wait, this seems to contradict your point #1 above.

On the other hand, it may mean that we can skip stage2 altogether, since 
we had stage2 during stage1 for 4.3...

Paolo

Re: 4.3 release plan

[Bernardo Innocenti]

Paolo Bonzini wrote:

>>> I got the first two, but what does the last one mean?
>> Wow, you're impressive...  The second one I thought would be too cryptic 
>> for all.
> 
> I got the second one too.  Italians must be good at acronyms. :-)

IMBGAA!

...BTTOWWTD!!!


>>> BTW, the tentative timeline says that 4.3 stage 1 will end 4 months 
>>> *ago*:
>>>   http://gcc.gnu.org/develop.html#timeline
>> Wait, this seems to contradict your point #1 above.

Uh?  On the countrary: saying that stage 1 should be finished
4 months ago *reinforces* the idea that 4.3 appears to be stable.


> On the other hand, it may mean that we can skip stage2 altogether, since 
> we had stage2 during stage1 for 4.3...

I'm definitely for short release cycles, so I'd generally
second this.  But in the wiki I see several stage1 projects
still waiting to be merged, as Steven Bosscher said

Would these have to go in now or later in 4.4?

-- 
   // Bernardo Innocenti
 \X/  http://www.codewiz.org/

Re: 4.3 release plan

[Paolo Bonzini]

>> I got the second one too.  Italians must be good at acronyms. :-)
> 
> IMBGAA!
> ...BTTOWWTD!!!

PBTMAICFOTL (Probably better than me as I cannot figure out the latter).

>> On the other hand, it may mean that we can skip stage2 altogether, since 
>> we had stage2 during stage1 for 4.3...
> 
> I'm definitely for short release cycles, so I'd generally
> second this.  But in the wiki I see several stage1 projects
> still waiting to be merged, as Steven Bosscher said
> 
> Would these have to go in now or later in 4.4?

I would propose waiting for these to be merged, and then moving directly 
to bug-fixing stage3.  We've effectively been in stage2 for 4 months.

Paolo

Re: 4.3 release plan

[Bernardo Innocenti]

Paolo Bonzini wrote:

>> ...BTTOWWTD!!!
> 
> PBTMAICFOTL (Probably better than me as I cannot figure out the latter).

Never mind, it was meant to be impossible to decode: "But The Third
One Was Way Too Difficult" :-)

>> Would these have to go in now or later in 4.4?
> 
> I would propose waiting for these to be merged, and then moving directly 
> to bug-fixing stage3.  We've effectively been in stage2 for 4 months.

I see, but what about all the pending stage 2 projects then?
These are most probably not yet ready to go in, and merging
too much stuff at the same time isn't a wise thing anyway.

(the next proposal is likely to cause some dissent)

What about moving 4.3 to stage 3 *now* and moving everything
else in 4.4 instead?  Hopefully, it will be a matter of just
a few months.  From http://gcc.gnu.org/gcc-4.3/changes.html,
it looks like it would already be quite a juicy release.

(ducks)

-- 
   // Bernardo Innocenti
 \X/  http://www.codewiz.org/

Re: 4.3 release plan

[mark]

On Sun, May 20, 2007 at 10:39:43PM -0700, Brooks Moses wrote:
> Bernardo Innocenti wrote:
> >(the next proposal is likely to cause some dissent)
> >What about moving 4.3 to stage 3 *now* and moving everything
> >else in 4.4 instead?  Hopefully, it will be a matter of just
> >a few months.  From http://gcc.gnu.org/gcc-4.3/changes.html,
> >it looks like it would already be quite a juicy release.
> Why?
> I mean, I suppose there could be advantages to doing this, but you 
> haven't mentioned even one.

I think a few people (me!) are waiting for GCJ to have official
support for Java 5 syntax and class libraries. Not that I would like
to rush you - or skip any valuable merges - but if the code that is in
right now is in a near ready state, waiting up to a year before
releasing seems unfortunate. :-(

Cheers,
mark

-- 
mark@mielke.cc / markm@ncf.ca / markm@nortel.com     __________________________
.  .  _  ._  . .   .__    .  . ._. .__ .   . . .__  | Neighbourhood Coder
|\/| |_| |_| |/    |_     |\/|  |  |_  |   |/  |_   | 
|  | | | | \ | \   |__ .  |  | .|. |__ |__ | \ |__  | Ottawa, Ontario, Canada

  One ring to rule them all, one ring to find them, one ring to bring them all
                       and in the darkness bind them...

                           http://mark.mielke.cc/

Re: 4.3 release plan

[Brooks Moses]

Bernardo Innocenti wrote:
> (the next proposal is likely to cause some dissent)
> 
> What about moving 4.3 to stage 3 *now* and moving everything
> else in 4.4 instead?  Hopefully, it will be a matter of just
> a few months.  From http://gcc.gnu.org/gcc-4.3/changes.html,
> it looks like it would already be quite a juicy release.

Why?

I mean, I suppose there could be advantages to doing this, but you 
haven't mentioned even one.

- Brooks

Re: 4.3 release plan

[Bernardo Innocenti]

Brooks Moses wrote:

>> What about moving 4.3 to stage 3 *now* and moving everything
>> else in 4.4 instead?  Hopefully, it will be a matter of just
>> a few months.  From http://gcc.gnu.org/gcc-4.3/changes.html,
>> it looks like it would already be quite a juicy release.
> 
> Why?
>
> I mean, I suppose there could be advantages to doing this, but you 
> haven't mentioned even one.

I apologize.  I guess you could get 10 different reasons
by 10 different people.

The reason _we_ care to get 4.3 sooner rather than later
is that we'd like to have the AMD Geode tuning and the
memcpy/strcpy() optimizations.

And also: why not?

-- 
   // Bernardo Innocenti
 \X/  http://www.codewiz.org/

Re: 4.3 release plan

[Andrew Pinski]

On 5/21/07, Bernardo Innocenti <bernie@codewiz.org> wrote:
> And also: why not?

I had hoped to get my pointer plus branch merged in which should
improve code gen and memory usage and compile time.

-- Pinski

Re: 4.3 release plan

[Joe Buck]

On Mon, May 21, 2007 at 11:31:19AM -0700, Andrew Pinski wrote:
> On 5/21/07, Bernardo Innocenti <bernie@codewiz.org> wrote:
> >And also: why not?
> 
> I had hoped to get my pointer plus branch merged in which should
> improve code gen and memory usage and compile time.

There seem to be quite a large number of not-yet-merged projects
on the wiki page at

http://gcc.gnu.org/wiki/GCC_4.3_Release_Planning

"Go straight to phase 3" would presumably mean that none of them get
in.  On the other hand, phase 1 has been running for a long time,
long past its original end date.  I'm surprised that phase 2 hasn't
started.

Re: 4.3 release plan

[Mike Stump]

On May 21, 2007, at 11:23 AM, Bernardo Innocenti wrote:
> The reason _we_ care to get 4.3 sooner rather than later
> is that we'd like to have the AMD Geode tuning

Submit to gcc 4.2.  Tuning seems to be the type of thing that should  
be safe to backport, if you really must have it.

Anyway, these reasons don't sound like reasons to goose the 4.3  
plan.  I'd argue for status quo.

Re: 4.3 release plan

[Bernardo Innocenti]

Joe Buck wrote:

>> I had hoped to get my pointer plus branch merged in which should
>> improve code gen and memory usage and compile time.
> 
> There seem to be quite a large number of not-yet-merged projects
> on the wiki page at

Never mind, I just did some investigation and it appears that
the patches we want are quite trivial to backport to 4.2, so
we don't need to rush 4.3 out.

> "Go straight to phase 3" would presumably mean that none of them get
> in.

But on the other end, if we do push GCC 4.3 out now, there
will be another stage 1 very soon :-)

Well, I don't want to step on anybody's feet. GCC has historically
had longer release cycles than other projects of comparable size
and this has clearly been consciously planned by experienced
people, most probably for good reason.

For the impatient, backporting is always an option.

-- 
   // Bernardo Innocenti
 \X/  http://www.codewiz.org/

Re: 4.3 release plan

[Bernardo Innocenti]

Mike Stump wrote:

> Submit to gcc 4.2.  Tuning seems to be the type of thing that should  
> be safe to backport, if you really must have it.

I extracted the relevant patches that would apply
to 4.2 as they were.  Currently regtesting just in
case.

-- 
   // Bernardo Innocenti
 \X/  http://www.codewiz.org/

Re: 4.3 release plan

[Bernardo Innocenti]

Bernardo Innocenti wrote:

> I extracted the relevant patches that would apply
> to 4.2 as they were.  Currently regtesting just in
> case.

Err, allow me to rephrase that more clearly: I have
extracted the Geode patches from the trunk and they
applied without modification to the 4.2 branch.

I'm currently bootstrapping 4.2 with "--with-arch=geode"
and will run the testsuite shortly.

-- 
   // Bernardo Innocenti
 \X/  http://www.codewiz.org/

Re: 4.3 release plan

[René Rebe]

On Monday 21 May 2007 20:23:46 Bernardo Innocenti wrote:
> Brooks Moses wrote:
> 
> >> What about moving 4.3 to stage 3 *now* and moving everything
> >> else in 4.4 instead?  Hopefully, it will be a matter of just
> >> a few months.  From http://gcc.gnu.org/gcc-4.3/changes.html,
> >> it looks like it would already be quite a juicy release.
> > 
> > Why?
> >
> > I mean, I suppose there could be advantages to doing this, but you 
> > haven't mentioned even one.
> 
> I apologize.  I guess you could get 10 different reasons
> by 10 different people.

12++

> The reason _we_ care to get 4.3 sooner rather than later
> is that we'd like to have the AMD Geode tuning and the
> memcpy/strcpy() optimizations.
> 
> And also: why not?

Stabelizing a major release series and then continuing to maintain
it is a time consuming efford. With a major series just started
introducing the next just some months later sound like wasting a
lot of our all time.

Backporting the Geode and *cpy tuning for just you local use-case
in contrast is not that hard at all.

Yours,

-- 
  René Rebe - ExactCODE GmbH - Europe, Germany, Berlin
  http://exactcode.de | http://t2-project.org | http://rene.rebe.name

Re: Regressions -march=geode/athlon regressions (Was: 4.3 release plan)

[Bernardo Innocenti]

Bernardo Innocenti wrote:

> I'm currently bootstrapping 4.2 with "--with-arch=geode"
> and will run the testsuite shortly.

It took a little longer than I expected because I was
seeing a few new regressions:

FAIL: gcc.c-torture/execute/20050316-2.c execution,  -O0
FAIL: gcc.c-torture/execute/20050316-2.c execution,  -O1
FAIL: gcc.c-torture/execute/20050316-2.c execution,  -O2
FAIL: gcc.c-torture/execute/20050604-1.c execution,  -O0
FAIL: gcc.c-torture/execute/20050604-1.c execution,  -O1
FAIL: gcc.c-torture/execute/20050604-1.c execution,  -O2
FAIL: gcc.c-torture/execute/20050604-1.c execution,  -O3 -fomit-frame-pointer
FAIL: gcc.c-torture/execute/20050604-1.c execution,  -O3 -fomit-frame-pointer -funroll-loops
FAIL: gcc.c-torture/execute/20050604-1.c execution,  -O3 -fomit-frame-pointer -funroll-all-loops -finline-functions
FAIL: gcc.c-torture/execute/20050604-1.c execution,  -O3 -g
FAIL: gcc.c-torture/execute/20050604-1.c execution,  -Os
FAIL: gcc.dg/compat/vector-2 c_compat_x_tst.o-c_compat_y_tst.o execute
FAIL: gcc.dg/builtin-apply4.c execution test

But then it turned out that using --with-arch=athlon produces
exactly the same regressions, while --with-arch=i586 is fine.

I analyzed only the first failing test.  It seems we fail
with all MMX enabled processors. But maybe the test has
invalid expectations, I'm not sure what it is trying to do.

 WORKS: gcc -O0 -march=i686   gcc.c-torture/execute/20050316-2.c && ./a.out
 FAILS: gcc -O0 -march=athlon gcc.c-torture/execute/20050316-2.c && ./a,out

This is with 4.2-branch, but it also happens with FC6's gcc-4.1.1.

-- 
   // Bernardo Innocenti
 \X/  http://www.codewiz.org/

Re: Regressions -march=geode/athlon regressions (Was: 4.3 release plan)

[Rask Ingemann Lambertsen]

On Tue, May 22, 2007 at 08:45:38PM -0400, Bernardo Innocenti wrote:
 
> FAIL: gcc.c-torture/execute/20050316-2.c execution,  -O0
[...]
> FAIL: gcc.c-torture/execute/20050604-1.c execution,  -O0
[...]
> 
> I analyzed only the first failing test.  It seems we fail
> with all MMX enabled processors. But maybe the test has
> invalid expectations, I'm not sure what it is trying to do.

   The last time I looked into it, the reason was use of MMX instructions
without (f)emms afterwards.

>  WORKS: gcc -O0 -march=i686   gcc.c-torture/execute/20050316-2.c && ./a.out
>  FAILS: gcc -O0 -march=athlon gcc.c-torture/execute/20050316-2.c && ./a,out
> 
> This is with 4.2-branch, but it also happens with FC6's gcc-4.1.1.

   4.3 is broken too. See e.g.
<URL:http://gcc.gnu.org/ml/gcc-testresults/2007-05/msg00635.html> which has
--with-arch=pentium3.

-- 
Rask Ingemann Lambertsen

Re: Regressions -march=geode/athlon regressions (Was: 4.3 release plan)

[Jakub Jelinek]

On Wed, May 23, 2007 at 10:40:46AM +0200, Rask Ingemann Lambertsen wrote:
> On Tue, May 22, 2007 at 08:45:38PM -0400, Bernardo Innocenti wrote:
>  
> > FAIL: gcc.c-torture/execute/20050316-2.c execution,  -O0
> [...]
> > FAIL: gcc.c-torture/execute/20050604-1.c execution,  -O0
> [...]
> > 
> > I analyzed only the first failing test.  It seems we fail
> > with all MMX enabled processors. But maybe the test has
> > invalid expectations, I'm not sure what it is trying to do.
> 
>    The last time I looked into it, the reason was use of MMX instructions
> without (f)emms afterwards.

Yeah:

        movq    -48(%ebp), %mm0 #, x
        call    test1   #
        movl    %eax, -40(%ebp) # D.1650, u.l
        movl    %edx, -36(%ebp) # D.1650, u.l
        flds    -40(%ebp)       # u.f
        flds    .LC0    #

(and test1 also using movq %mm0, something, also no (f)emms).

	Jakub

Re: Regressions -march=geode/athlon regressions (Was: 4.3 release plan)

[Bernardo Innocenti]

On Wed, 23 May 2007 10:49:53 +0200, Jakub Jelinek <jakub@redhat.com> wrote:

>>    The last time I looked into it, the reason was use of MMX
>> instructions without (f)emms afterwards.
> 
> Yeah:
> [...]
> (and test1 also using movq %mm0, something, also no (f)emms).

Is it already filed somewhere in Bugzilla?  Or would you
like me to open a new bug for it?

-- 
  // Bernardo Innocenti
\\X/  http://www.codewiz.org/

Re: Regressions -march=geode/athlon regressions (Was: 4.3 release ?plan)

[Rask Ingemann Lambertsen]

On Wed, May 23, 2007 at 10:57:38AM +0200, Bernardo Innocenti wrote:
> 
> On Wed, 23 May 2007 10:49:53 +0200, Jakub Jelinek <jakub@redhat.com> wrote:
> 
> >>    The last time I looked into it, the reason was use of MMX
> >> instructions without (f)emms afterwards.
> > 
> > Yeah:
> > [...]
> > (and test1 also using movq %mm0, something, also no (f)emms).
> 
> Is it already filed somewhere in Bugzilla?  Or would you
> like me to open a new bug for it?

   <URL:http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19161> but nobody seems
to be working on it.
 
-- 
Rask Ingemann Lambertsen