This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug tree-optimization/85416] New: Massive performance regression when switching on "-march=native"


https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85416

            Bug ID: 85416
           Summary: Massive performance regression when switching on
                    "-march=native"
           Product: gcc
           Version: 8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: martin@mpa-garching.mpg.de
  Target Milestone: ---

Created attachment 43945
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=43945&action=edit
Test case

When compiling the attached test case with

g++ -O3 -std=c++14 testcase.cc

and measuring the runtime via

time ./a.out

I get a CPU time of roughly 1.3 seconds

Recompiling with

g++ -O3 -std=c++14 -march=native testcase.cc

increases run time to over 8 seconds.

CPU: Intel Core i5-5300U

This was measured on trunk (gcc version 8.0.1 20180411), but I see similar
performance degradations already in gcc 5.4.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]