On 02/16/2012 11:08 AM, Gregory Maxwell wrote:
On Thu, Feb 16, 2012 at 10:25 AM, Vladimir Makarov<vmakarov@xxxxxxxxxx> wrote:
GCC has a big community of very dedicated people. LLVM has no such
community. So IMHO GCC will be more high quality compiler than LLVM until
LLVM gets such community.
That can't be expected to continue now that there are many employers
hiring people and forbidding them from working on GCC, even in their
own time, while permitting them to work on LLVM.
That is why I wrote 'until'. Apple has a lot of resources (e.g. i
guess they are just buying BSD developers to try them to compile the
system by LLVM). There is some danger for GCC community from LLVM.
I'd say now companies hiring to work on LLVM and GCC is about 50/50%
(that is my impression of checking http://compilerjobs.com/ for long
time). Also LLVM is more popular for younger people because of steep
learning curve. GCC has more experienced compiler specialists. LLVM
community simply can overlive GCC community if GCC community has not
enough fresh blood.
But still LLVM are many years behind GCC generated code quality
(although for many programs better quality is not important) and it is
much harder to get 10-15% improvement (see Proebsting's law) especially
when you already have classical optimizations implemented: some people
says with some point it needs exponential efforts.
But the most important is differences in views of Apple and open source
projects (like Fedora) on software role and development in the world.
When Apple used GCC they were quit bad at contributing their code to
community. It could happen LLVM the same if they have such choice.
I'm not just spreading sour news for the sake of it, here is an
example of where I ran into this impeding a GCC crash bug:
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50335#c8
I checked the bugzilla. The graphyte project is pain for GCC community
too. It promised a lot but never delivered it. Most developers were
from AMD and AMD finally fired all of them. So we have practically no
people working on it. Also their approach had some mistakes even with
the design point of view, some bugs will never be fixed because of
underlying used libraries (which are not part of GCC project).
(Though it will be quite ironic when LLVM becomes unusable to everyone
because the "we don't give up our patent rights for this when we
contribute to it" turns it into a thicket)
--
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/devel
