[llvm-dev] Ineffective code after loop unrolling with -O3, ok with -Os

[Dávid Bolvanský via llvm-dev]

Fixed link: https://godbolt.org/g/2EqCKK

2018-04-03 18:15 GMT+02:00 Dávid Bolvanský <david.bolvansky at gmail.com>:

> Hi all,
>
> I found some issues during my testing of "loop unrolling" capabilities of
> LLVM's opt.
>
> Seems like LLVM generates slower code with -O3 since it wrongly decides to
> unroll a simple loop.
> With option -Os, no loop unrolling, the output looks well.
>
> Code:
> https://godbolt.org/#g:!((g:!((g:!((h:codeEditor,i:(j:1,
> lang:c%2B%2B,source:'%23include+%3Ccstdio%3E%0A%
> 0Aint+f(void)%0A%7B%0A++int+n+%3D+100%3B%0A++while+(--n)+%
> 7B%0A++++++puts(%22a%22)%3B%0A++%7D%0A++return+0%3B%0A%7D%
> 0A%0Aint+main(void)+%7B%0A++++f()%3B%0A++++return+0%3B%0A%7D'),l:'5
> ',n:'0',o:'C%2B%2B+source+%231',t:'0')),k:33.
> 61947345592715,l:'4',n:'0',o:'',s:0,t:'0'),(g:!((h:compiler,
> i:(compiler:g73,filters:(b:'0',binary:'1',commentOnly:'0',
> demangle:'0',directives:'0',execute:'1',intel:'0',trim:'0'
> ),lang:c%2B%2B,libs:!(),options:'-O3',source:1),l:'5',
> n:'0',o:'x86-64+gcc+7.3+(Editor+%231,+Compiler+%231)+C%
> 2B%2B',t:'0')),header:(),k:33.04719321073953,l:'4',n:'0',o:'
> ',s:0,t:'0'),(g:!((h:compiler,i:(compiler:clang600,filters:(
> b:'0',binary:'1',commentOnly:'0',demangle:'0',directives:'0'
> ,execute:'1',intel:'0',trim:'0'),lang:c%2B%2B,libs:!(),
> options:'-O3',source:1),l:'5',n:'0',o:'x86-64+clang+6.0.0+(
> Editor+%231,+Compiler+%232)+C%2B%2B',t:'0')),k:33.
> 33333333333333,l:'4',n:'0',o:'',s:0,t:'0')),l:'2',n:'0',o:''
> ,t:'0')),version:4
>
> Thanks
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20180403/462ea823/attachment.html>