Re: C++ interop issue due to non-null pointers

2014-06-18 Thread Florian Weimer
On 06/17/2014 05:00 PM, Jakub Jelinek wrote: GCC will likely not optimize it away at this point, but having code with undefined behavior is just asking for future trouble. Just use "" instead? It's always const and may lack sufficient alignment. The former isn't a problem in C++ (I think), b

Re: C++ interop issue due to non-null pointers

2014-06-18 Thread Jakub Jelinek
On Wed, Jun 18, 2014 at 09:18:10AM +0200, Florian Weimer wrote: > On 06/17/2014 05:00 PM, Jakub Jelinek wrote: > > >>>GCC will likely not optimize it away at this point, but having code with > >>>undefined behavior is just asking for future trouble. Just use "" instead? > >> > >>It's always const

Re: C++ interop issue due to non-null pointers

2014-06-18 Thread Florian Weimer
On 06/18/2014 09:24 AM, Jakub Jelinek wrote: You're returning a T *, not a void *, and C++ requires that pointers are properly aligned even if they aren't dereferenced. C qsort doesn't return anything and the comparison function returns int. extern void qsort (void *__base, size_t __nmemb, si

Re: [GSoC] decision tree first steps

2014-06-18 Thread Prathamesh Kulkarni
On Tue, Jun 17, 2014 at 3:15 PM, Richard Biener wrote: > > On Tue, Jun 17, 2014 at 12:21 AM, Prathamesh Kulkarni > wrote: > > On Mon, Jun 16, 2014 at 4:45 PM, Richard Biener > > wrote: > >> > >> > * Patterns requiring GENERIC support like cond_expr > >> > I am not sure about how to handle these

Re: [GSoC] decision tree first steps

2014-06-18 Thread Richard Biener
On Wed, Jun 18, 2014 at 11:21 AM, Prathamesh Kulkarni wrote: > On Tue, Jun 17, 2014 at 3:15 PM, Richard Biener > wrote: >> >> On Tue, Jun 17, 2014 at 12:21 AM, Prathamesh Kulkarni >> wrote: >> > On Mon, Jun 16, 2014 at 4:45 PM, Richard Biener >> > wrote: >> >> >> >> > * Patterns requiring GENER

Re: regs_used estimation in IVOPTS seriously flawed

2014-06-18 Thread Richard Biener
On Tue, Jun 17, 2014 at 4:59 PM, Bingfeng Mei wrote: > Hi, > I am looking at a performance regression in our code. A big loop produces > and uses a lot of temporary variables inside the loop body. The problem > appears that IVOPTS pass creates even more induction variables (from original > 2 to 27

RE: regs_used estimation in IVOPTS seriously flawed

2014-06-18 Thread Bingfeng Mei
> -Original Message- > From: Richard Biener [mailto:richard.guent...@gmail.com] > Sent: 18 June 2014 12:36 > To: Bingfeng Mei > Cc: gcc@gcc.gnu.org > Subject: Re: regs_used estimation in IVOPTS seriously flawed > > On Tue, Jun 17, 2014 at 4:59 PM, Bingfeng Mei wrote: > > Hi, > > I am lo

Re: [GSoC] How to get started with the isl code generation

2014-06-18 Thread Roman Gareev
Hi Tobias, I made a separate patch and rebased the previous one. They are attached to this letter. > I am surprised. Are all these includes really needed to get _this_ patch > compile? (I asked this before). I saw your previous comment related to this and the following includes were removed: is

[GSoC] question about regular expressions

2014-06-18 Thread Roman Gareev
Dear gcc contributors, could you please advise how to better write the following testcase? After the compilation with -O2 -fdump-tree-graphite-all -fgraphite-identity -fgraphite-code-generator=isl the dump file should contain the following text ISL AST generated by ISL: for (int c1 = 0; c1 < n -

Re: [GSoC] How to get started with the isl code generation

2014-06-18 Thread Tobias Grosser
On 18/06/2014 15:22, Roman Gareev wrote: Hi Tobias, I made a separate patch and rebased the previous one. They are attached to this letter. I am surprised. Are all these includes really needed to get _this_ patch compile? (I asked this before). I saw your previous comment related to this an

gfortran option -fno-automatic and -finit-local-zero

2014-06-18 Thread Chase, Jeffrey G. (IS) (Contr)
Hello, First of all thanks so much for all your hard work!  I'm compiling a bunch of legacy fortran that relies on SAVE semantics from relic compilers.  I'm using the -fno-automatic option to replicate that behavior.  I also was using the -finit-local-zero option, assuming that it would initial

Re: [GSoC] How to get started with the isl code generation

2014-06-18 Thread Roman Gareev
I used trunk and compiled these patches only with isl 0.12 and ClooG 0.18.1. Which versions of these libraries are need to be checked for compatibility? -- Cheers, Roman Gareev

Re: [GSoC] How to get started with the isl code generation

2014-06-18 Thread Tobias Grosser
On Jun 18, 2014 6:03 PM, Roman Gareev wrote: > > I used trunk and compiled these patches only with isl 0.12 and ClooG > 0.18.1. Which versions of these libraries are need to be checked for > compatibility? That's fine. Please post the patches for wider review at gcc patches. Also mention that

ICE with atomic_store

2014-06-18 Thread Basile Starynkevitch
Hello All, The following code: #include struct s1_st { char* i_name; struct s1_st* i_foo; }; void clear_s1 (struct s1_st*s) { __atomic_store(s->i_name, NULL, __ATOMIC_SEQ_CST); } gives an ICE when compiled (on Debian/Sid/amd64) by GCC 4.8 & 4.9 (I did not

Re: ICE with atomic_store

2014-06-18 Thread Andrew MacLeod
On 06/18/2014 12:59 PM, Basile Starynkevitch wrote: Hello All, The following code: #include struct s1_st { char* i_name; struct s1_st* i_foo; }; void clear_s1 (struct s1_st*s) { __atomic_store(s->i_name, NULL, __ATOMIC_SEQ_CST); } gives an ICE when c

gcc-4.9-20140618 is now available

2014-06-18 Thread gccadmin
Snapshot gcc-4.9-20140618 is now available on ftp://gcc.gnu.org/pub/gcc/snapshots/4.9-20140618/ and on various mirrors, see http://gcc.gnu.org/mirrors.html for details. This snapshot has been generated from the GCC 4.9 SVN branch with the following options: svn://gcc.gnu.org/svn/gcc/branches