gcc-patches-ow...@gcc.gnu.org wrote on 28/12/2011 11:05:19 PM:
> Hi,
Hi Igor,
>
> Here is another patch about failures in gcc.dg/vect tests. These
> changes fix fails that could be seen on avx-built compilers. It also
> introduces no FAILs/XFAILs/XPASSes/ERRORs on regular i686, x86_64,
> avx2_
Hello world,
the attached patch fixes PR 51502, where we wrongly recognized a
procedure as implicit pure when we were assigning to a module
variable within a block. This is a potential cause for
wrong-code regressions (although no actual test case
exists).
For the test case, I had to scan for t
Ping.
Ayal, could you review this patch and these two patches too.
http://gcc.gnu.org/ml/gcc-patches/2011-12/msg00505.html
http://gcc.gnu.org/ml/gcc-patches/2011-12/msg00506.html
Happy holidays.
2011/12/7 Roman Zhuykov :
> Apologies for the messed up previous e-mail.
>
> 2011/10/12 Ayal Zaks :
>>
Because it includes AVX and AVX2 which deals with int and for AVX2
there are no problems with doubled diagnostics.
I understand that all this looks quite bulky but it's hard to create
something which looks better without loosing generality
On Thu, Dec 29, 2011 at 4:15 PM, Ira Rosen wrote:
>
>
>
Igor Zamyatin wrote on 29/12/2011 02:04:45 PM:
> When compiler configured with, say corei7-avx, it outputs twice more
> diagnostics on integer tests since AVX deals mostly with floats. I.e.
> compiler tries to vectorize on AVX vector size, than fails and then
> vectorizes on smaller vector size
Several of the gcc.dg tests were failing on 32-bit powerpc-linux since the
tests test the machine independent vector support, and the powerpc backend
issues a warning on passing vectors if -mabi=altivec was not used (it is on by
default for 64-bit). I committed the following patch as obvious to ad
Igor Zamyatin wrote on 29/12/2011 02:29:46 PM:
>
> Because it includes AVX and AVX2 which deals with int and for AVX2
> there are no problems with doubled diagnostics.
And you can't just update vect_int because AVX does support it but with
128-bit vectors, right?
So, your vect_float_no_int loo
This week I investigated modulo scheduler on IA64. Enabling SMS by default
(-fmodulo-sched -fmodulo-sched-allow-regmoves) leads to bootstrap failure
on IA64: gcc/build/genautomata.o differs while comparing stages 2 and 3.
I haven't studied this issue in detail, because the combination of these
my
PR libstdc++/51701
* doc/xml/manual/extensions.xml (Input and Output): Remove reference
to RWLock class.
Committed to trunk.
Index: doc/xml/manual/extensions.xml
===
--- doc/xml/manual/extensions.xml (rev
On Tue, Dec 13, 2011 at 12:35 PM, Richard Henderson wrote:
> On 12/12/2011 07:08 PM, David Edelsohn wrote:
>> If you don't want to grab the L2 cache line size directly, could you
>> default to 32 bytes on PPC32 and 128 bytes on PPC64 (__powerpc64__) ?
>
> Done.
Richard,
By the way, the patch is
Thanks, I'll look into your remarks
On Thu, Dec 29, 2011 at 5:33 PM, Ira Rosen wrote:
>
>
> Igor Zamyatin wrote on 29/12/2011 02:29:46 PM:
>
>>
>> Because it includes AVX and AVX2 which deals with int and for AVX2
>> there are no problems with doubled diagnostics.
>
> And you can't just update v
Ilya is on vacation so I'll make the answer.
Overall score became worse on 0.3%.
> -Original Message-
> From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-ow...@gcc.gnu.org] On
> Behalf Of Vladimir Makarov
> Sent: Thursday, December 22, 2011 8:15 PM
> To: Ilya Enkovich
> Cc: gcc-pat
On 12/29/2011 06:59 AM, Torvald Riegel wrote:
On Wed, 2011-12-28 at 21:03 -0500, Patrick Marlier wrote:
With i386, the regparm(2) is not taken into account when there is a
variadic function. All parameters are in the stack.
Since this variable argument is never used removing it is not a problem.
This patch makes MicroBlaze handle return and simple_return in
the same fashion as MIPS.
2011-12-29 Michael Eager
* config/microblaze/microblaze.md: Add expander for simple_return,
return, add return_internal and simple_return_internal insns.
Checked in, revision 182725.
--
When compiler configured with, say corei7-avx, it outputs twice more
diagnostics on integer tests since AVX deals mostly with floats. I.e.
compiler tries to vectorize on AVX vector size, than fails and then
vectorizes on smaller vector size. This double work leads to double
diagnostic output.
On T
I committed the following patch as obvious to fix some testsuite failures on
the PowerPC SPE. I did verify that they don't cause any regressions on
powerpc-linux.
2011-12-29 Michael Meissner
PR testsuite/50988
* gcc.target/powerpc/altivec-34.c: Require altivec support.
On Wed, 2011-12-28 at 21:03 -0500, Patrick Marlier wrote:
> With i386, the regparm(2) is not taken into account when there is a
> variadic function. All parameters are in the stack.
> Since this variable argument is never used removing it is not a problem.
The ABI specifies beginTransaction as a
On Thu, 2011-12-29 at 19:00 +0100, François Dumont wrote:
> Attached patch applied.
>
> 2011-12-29 François Dumont
>
> PR libstdc++/51608
> [...]
After this patch I'm getting the following error when building a gcc
cross compiler with "make all":
/home/yam/code/gcc/gcc-trunk
On Thu, 29 Dec 2011, Oleg Endo wrote:
On Thu, 2011-12-29 at 19:00 +0100, François Dumont wrote:
Attached patch applied.
2011-12-29 François Dumont
PR libstdc++/51608
[...]
After this patch I'm getting the following error when building a gcc
cross compiler with "make all"
On Wed, Dec 28, 2011 at 12:34:25PM -0800, Richard Henderson wrote:
> On 12/28/2011 09:39 AM, Michael Meissner wrote:
> > && in_section != text_section
> > - && !unlikely_text_section_p (in_section)
> > + && (in_section && (in_section->common.flags & SECTION_CODE)) == 0
>
> You should
On Thu, 2011-12-29 at 21:51 +0100, Marc Glisse wrote:
> On Thu, 29 Dec 2011, Oleg Endo wrote:
>
> > On Thu, 2011-12-29 at 19:00 +0100, François Dumont wrote:
> >> Attached patch applied.
> >>
> >> 2011-12-29 François Dumont
> >>
> >> PR libstdc++/51608
> >> [...]
> >
> > After
On 12/29/2011 10:33 PM, Oleg Endo wrote:
Yes, _N is defined somewhere else (no idea where yet...). Renaming _N
to _Nn helps.
_N in any case is a Solaris (at least) badname, should never be used.
Likewise any other _ single capital, to be safe. I'll fix that.
Francois, please be more careful in
... applied the below (sanity checked x86_64-linux).
Thanks,
Paolo.
///
2011-12-29 Paolo Carlini
* include/bits/hashtable_policy.h (struct _Ebo_helper<>): Don't use
_N, badname on Solaris; minor stylistic changes.
Index: include/bits/hashtable_policy.h
===
Thank you Mike. Will make the change before commit.
On 12/28/11 8:53 PM, Mike Stump wrote:
On Dec 28, 2011, at 1:52 PM, Nenad Vukicevic wrote:
-systems (Leopard 10.5.7+ and Snow Leopard 10.6)
+systems (Leopard 10.5.7+, Snow Leopard 10.6, and Lion 1.7)
1.7? Should this be 10.7?
Hello,
This patch add a static branch predict heuristic for the following cases:
for (int i = 0; i < bound; i++) {
if (i < bound - 2)
computation_1();
else
computation_2();
}
In this case, we would predict the branch to be taken because it's
comparing loop induction variable to loop
The audit trail in the PR pretty much sums up the problem.
OK to commit?
2011-12-29 Steven G. Kargl
* inclhack.def: Disgusting hack to workaround brain damage of
defining __cplusplus as 201103L with -std=c++11 when g++ does
not support c++11.
* fixincl.x: regen
PING 8: For reviewing the modified CR16 port.
Hello,
Can some one please review the updated patch and let me know if any more
changes are required to be done in it?
Rainer had suggested few important changes last time. After making those
changes, the modified patch was posted at following URL:
27 matches
Mail list logo
