This is the mail archive of the
mailing list for the GCC project.
Re: RFC: Update ISL under gcc/infrastructure/ ? // Remove CLooG?
- From: Richard Biener <richard dot guenther at gmail dot com>
- To: Tobias Grosser <tobias at grosser dot es>
- Cc: Sven Verdoolaege <skimo at kotnet dot org>, Mircea Namolaru <mircea dot namolaru at inria dot fr>, GCC Development <gcc at gcc dot gnu dot org>
- Date: Thu, 6 Nov 2014 11:15:19 +0100
- Subject: Re: RFC: Update ISL under gcc/infrastructure/ ? // Remove CLooG?
- Authentication-results: sourceware.org; auth=none
- References: <20141104151755 dot GA13200 at physik dot fu-berlin dot de> <545AB53E dot 5080808 at grosser dot es> <CABGF_gfmcazQLXPxZ4ET3P0VCNPYxYazJGWDY3GsGzFaj1ueGA at mail dot gmail dot com> <545B1D00 dot 6060605 at grosser dot es> <CAFiYyc0aBCDDdgQPvDszdzG7kk6RRnKMWixtUkDr10T-PnMiXw at mail dot gmail dot com> <545B42B4 dot 4070701 at grosser dot es>
On 11/6/14, Tobias Grosser <email@example.com> wrote:
> On 06.11.2014 10:05, Richard Biener wrote:
>> On Thu, Nov 6, 2014 at 8:02 AM, Tobias Grosser <firstname.lastname@example.org> wrote:
>>> On 06.11.2014 07:04, Roman Gareev wrote:
>>>>> CLooG is not necessarily needed. You can run graphite just with ISL.
>>>>> main reason that ISL code generation is not enabled by default is that
>>>>> did not yet get extensive testing and it was unclear who will have the
>>>>> to fix possible bugs.
>>>> Could you please advise me which test suites should be used to make
>>>> performance comparison between CLooG and ISL generator? (I would like
>>>> to do this, even though the old generator is removed).
>>> I do not have specific advices. You can use various open source
>>> e.g. the LLVM test suite or, if you have access, you could run SPEC or
>>>>> @Mircae, Roman: Would you have time to help with bug-fixing if we do
>>>>> switch now? (I am happy to review patches and give advice, but can not
>>>>> the full move myself)
>>>> I could find time for this. What do you mean by âswitchâ? If Iâm not
>>>> mistaken, ISL generator is already used by default. Should we remove
>>>> support of CLooG generator and all files related to it?
>>> Wow, I must really have been sleeping (or just forgetting). The switch
>>> already happened. This is amazing.
>>> As the ISL code generator has been default since a while and we did not
>>> many bug reports, the actual switch seems to have worked well. We could
>>> probably still need some testing, but in this case it is most likely time
>>> drop the CLooG support entirely. Are you interested to provide the
>>> Also, as Tobias suggested we should raise the minimal supported isl level
>>> 0.14 to be sure PR 62289 is fixed.
>> As I am testing with system isl/cloog that would be unfortunate as it
>> I'd either drop testing graphite for 4.8 and 4.9 or for 5.0 ... AFAIK
>> 12.x and 13+ cannot co-exist in the system due to include file conflicts.
> is there any chance we can add the deprecated isl_int includes back into
> isl 0.14.1. This would unblock the testing and we could remove them as
> soon as gcc 4.8/4.9 has been phased out.
> This would also fix my polly code coverage tests which do not work since
> isl_int has been moved into different header files, as ubuntu does not
> want to update isl as long as such an update breaks gcc.
Ah, so it's still there? Maybe we can simply add some configury to detect
its location and fix build with a patch for 4.8 and 4.9? Is there maybe even
a preprocessor macro one can check that newer ISL provide that can tell
us where to look for isl_int?