[Bug target/95784] Failure to optimize usage of __builtin_add_overflow with return statement properly

[gabravier at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95784

Gabriel Ravier <gabravier at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #48760|0                           |1
        is obsolete|                            |
  Attachment #48761|0                           |1
        is obsolete|                            |

--- Comment #8 from Gabriel Ravier <gabravier at gmail dot com> ---
Created attachment 48763
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=48763&action=edit
File for benchmarking this function but everything is aligned properly and the
result is used

I've added another benchmark that uses the result of the operation (stores it
on the stack). Here are the average time taken to execute the program over 10
runs of each version :

GCC's version : 2.418 seconds
LLVM's version : 2.160 seconds
The "do nothing" version : 2.107 seconds

Do tell me if the benchmark is still flawed.