[Bug target/91719] gcc compiles seq_cst store on x86-64 differently from clang/icc

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91719

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Tried:
unsigned int a;

int
main ()
{
  for (int i = 0; i < 100000000; i++)
    __atomic_store_n (&a, i, __ATOMIC_SEQ_CST);
  return 0;
}
and:
unsigned int a;

int
main ()
{
  for (int i = 0; i < 100000000; i++)
    __atomic_exchange_n (&a, i, __ATOMIC_SEQ_CST);
  return 0;
}
and got (microbenchmark, sure):
Intel(R) Core(TM) i9-7960X CPU @ 2.80GHz
user    0m1.045s
user    0m0.441s
Intel(R) Xeon(R) CPU E5-2697 v3 @ 2.60GHz
user    0m1.216s
user    0m0.529s
Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz
user    0m1.185s
user    0m0.627s
Intel Xeon E312xx (Sandy Bridge)
user    0m1.600s
user    0m0.846s
Quad-Core AMD Opteron(tm) Processor 8354
user    0m1.720s
user    0m0.724s