Page 1 of 2 12 LastLast
Results 1 to 10 of 21

Thread: LLVM/Clang 3.2 Compiler Competing With GCC

Hybrid View

  1. #1
    Join Date
    Jan 2007
    Posts
    15,108

    Default LLVM/Clang 3.2 Compiler Competing With GCC

    Phoronix: LLVM/Clang 3.2 Compiler Competing With GCC

    With last week's release of LLVM 3.2, here are new benchmarks of LLVM 3.2 with the Clang C/C++ compiler front-end. The LLVM/Clang 3.2 performance using last week's source code releases were compared to the earlier LLVM/Clang 3.1 release and then for competition was the GCC 4.7.2 stable release and the latest GCC 4.8.0 development snapshot.

    http://www.phoronix.com/vr.php?view=18321

  2. #2
    Join Date
    Sep 2012
    Posts
    289

    Default GCC Wins hands down

    GCC wins hands down. GO and cry Apple fanboys !!!

  3. #3
    Join Date
    May 2011
    Posts
    1,558

    Default

    Does anyone know how the two compare on ARM processors? This seems to be an area where the two compete head-to-head.

  4. #4
    Join Date
    Oct 2009
    Posts
    845

    Default

    Quote Originally Posted by johnc View Post
    Does anyone know how the two compare on ARM processors? This seems to be an area where the two compete head-to-head.
    Yes this is an area which I'm also interested in, in x86 64 which I am benchmarking GCC still holds supreme in terms of code performance, however I have no idea of how these compilers compare on ARM.

    And given that this is an extremely strong architecture outside of the desktop it would be very interesting to see which compiler toolchain has the best performance here.

  5. #5
    Join Date
    Aug 2012
    Location
    Pennsylvania, United States
    Posts
    1,911

    Default

    Quote Originally Posted by wargames View Post
    GCC wins hands down. GO and cry Apple fanboys !!!
    A) Just supporting LLVM doesn't make someone an "apple fanboy" so go and take your apparent GCC-fanboyism somewhere else.

    B) "Hands down" is a BIG stretch. LLVM was extremely competitive in just about every test that doesn't require OpenMP, and once OpenMP IS sorted out I bet those benchmarks will show LLVM is about the same as GCC in those tests as well.

    C) The benefits in LLVM aren't in speed. 1) Modularity-- something GCC cant accomplish without scrapping a large chunk of the codebase SPECIFICALLY because GNU WANTED a monolithic codebase so that companies couldn't pull in select portions of gcc and just use those. 2) Cleaner code base. 3) Competition finally amongst compilers finally. GCC has gone uncontested in the open source world for a very long time, its good to see a project finally giving it a run for its money and ensuring that they don't start to rest on their laurels

  6. #6
    Join Date
    Oct 2012
    Posts
    293

    Default

    Quote Originally Posted by Ericg View Post
    A) Just supporting LLVM doesn't make someone an "apple fanboy" so go and take your apparent GCC-fanboyism somewhere else.

    B) "Hands down" is a BIG stretch. LLVM was extremely competitive in just about every test that doesn't require OpenMP, and once OpenMP IS sorted out I bet those benchmarks will show LLVM is about the same as GCC in those tests as well.

    C) The benefits in LLVM aren't in speed. 1) Modularity-- something GCC cant accomplish without scrapping a large chunk of the codebase SPECIFICALLY because GNU WANTED a monolithic codebase so that companies couldn't pull in select portions of gcc and just use those. 2) Cleaner code base. 3) Competition finally amongst compilers finally. GCC has gone uncontested in the open source world for a very long time, its good to see a project finally giving it a run for its money and ensuring that they don't start to rest on their laurels
    to 1) and this was and still is a viable reason
    to 2) can't judge... don't know it and do not care neither.
    to 3) well, as the test results show they did quite well over a lot of years even without competition. actually i think this shows quite well one major plus of oos, they do not really need competition to be pushed forward. imagine how bad a closed source compiler would have evolved without any competion? the few devs wouldn't care about issues...

    edit: before somebody gets me wrong: i do wellcome the new competition.

  7. #7
    Join Date
    Jun 2011
    Posts
    1,030

    Default

    Quote Originally Posted by a user View Post
    to 1) and this was and still is a viable reason
    to 2) can't judge... don't know it and do not care neither.
    to 3) well, as the test results show they did quite well over a lot of years even without competition. actually i think this shows quite well one major plus of oos, they do not really need competition to be pushed forward. imagine how bad a closed source compiler would have evolved without any competion? the few devs wouldn't care about issues...

    edit: before somebody gets me wrong: i do wellcome the new competition.
    No even Open Source Software absolutely needs competition, particularly when it's a monolithic product as opposed to a more modular one*. The only reason that GCC is somewhat faster at this point in time is that it's been being developed since 1987, LLVM on the other hand was started in 2003, which means GCC has 13 years of development time in advance of LLVM, that's more than twice the lifespan of LLVM.

    A perfect example of the competition bringing forth better products in OSS here with GCC being that GCC now has decent error reporting capability because of the LLVM people having done it. As well as a relatively large refactoring of the GCC codebase into C++ and otherwise.

    What we're seeing here though is the natural result in the difference in design between Modular and Monolithic architectures, a modular project make take a bit longer for initial design and setting up but once it's there it can accelerate far faster and far longer (almost if not indefinitely) whereas a monolithic application is much quicker to design and set up but without refactoring to a more modular design the development will slow eventually to a halt no matter how many people you throw at it, just because the codebase is so unwieldy, thus allowing a properly modularly designed application to accelerate past it.

    * Now theoretically a Completely Modular codebase could exist in a vaccuum and do quite well, but it still benefits if nothing else from pulling ideas from competing products

  8. #8
    Join Date
    Oct 2012
    Posts
    293

    Default

    Quote Originally Posted by Luke_Wolf View Post
    No even Open Source Software absolutely needs competition,
    why "no"? did i claimed the opposite? no, i didn't! i just pointed out that in this case it worked well even without AND actually oos without competition has better chances to eveolve than closed source without competition.

    this does not mean that competition in general is also for oos recommended.

    particularly when it's a monolithic product as opposed to a more modular one*. The only reason that GCC is somewhat faster at this point in time is that it's been being developed since 1987, LLVM on the other hand was started in 2003, which means GCC has 13 years of development time in advance of LLVM, that's more than twice the lifespan of LLVM.
    no, this is wrong! actually this has quite little to do with gcc's current state of performance. if your claim would be true than gcc wouldn't had performed so well in comparision with other compilers that existed over the years. just consider ms c++ compiler which actually performes quite bad, or compare it to the intel compiler, which used to perform better, though gcc still could hold up quite well.

    the reason why llvm start to enhance so fast is that it could and did learned from the past of the others.

    A perfect example of the competition bringing forth better products...
    actually you don't need to post this here as i haven't seen somebody denieng that competition is good.

    What we're seeing here though is the natural result in the difference in design between Modular and Monolithic architectures, a modular project make take a bit longer for initial design and setting up but once it's there it can accelerate far faster and far longer (almost if not indefinitely) whereas a monolithic application is much quicker to design and set up but without refactoring to a more modular design the development will slow eventually to a halt no matter how many people you throw at it, just because the codebase is so unwieldy, thus allowing a properly modularly designed application to accelerate past it.
    you are simplifying things quite a lot. it is easy for interpret things like this while ignoring the fact that llvm can look back on the work of all the other compiler devs and learn from it.

    * Now theoretically a Completely Modular codebase could exist in a vaccuum and do quite well, but it still benefits if nothing else from pulling ideas from competing products
    i like how you are talking about a "completely modular codebase". there is no completely modular codebase, and even less a viable definition of such a thing. you can overmodularize things and make it worse. there is always the need of a design decision what and how much do you modularize. and afterwards you always see you could have done it better. LLVM has the advantage to look "afterwards".

    this helps llvm to improve fast to the current level of gcc. but if it can top it significantly is yet to be seen.

  9. #9
    Join Date
    Nov 2007
    Posts
    1,024

    Default

    Quote Originally Posted by a user View Post
    to 1) and this was and still is a viable reason
    to 2) can't judge... don't know it and do not care neither.
    to 3) well, as the test results show they did quite well over a lot of years even without competition. actually i think this shows quite well one major plus of oos, they do not really need competition to be pushed forward. imagine how bad a closed source compiler would have evolved without any competion? the few devs wouldn't care about issues...

    edit: before somebody gets me wrong: i do wellcome the new competition.
    GCC had major competition, what are you talking about? It's still the unpopular choice in a number of important fields where other compilers are dominant. Even in fields where it is dominant, the versions used are often forks of ancient crufty versions (in some cases 3.x, not even 4.x) with tons of vendor-specific changes (and the GPL is barely useful here since the FSF requires copyright assignment to roll code back into GCC proper, which said vendors have zero interest in granting, and hobbyist developers have little interest in patching, fixing, forking, or maintaining ancient vendor forks).

  10. #10
    Join Date
    Mar 2008
    Posts
    575

    Default

    I think it is safe to say that GCC has finally a competitor.

    Sure, they are released under different licences BUT we all agree that variety is man's wealth, do we?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •