[Bug bootstrap/63933] New: Build stage1 with -O2 during bootstrap if host compiler is a recent gcc version

[Bug bootstrap/63933] New: Build stage1 with -O2 during bootstrap if host compiler is a recent gcc version

[trippels at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63933

            Bug ID: 63933
           Summary: Build stage1 with -O2 during bootstrap if host
                    compiler is a recent gcc version
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: bootstrap
          Assignee: unassigned at gcc dot gnu.org
          Reporter: trippels at gcc dot gnu.org

Currently gcc doesn't add any optimization level flags during stage1.
This means that stage1 is build with the default -O0 and that results
in a very slow compiler. Build times during stage2 are very high
because of this.

One solution would be to detect the host compiler and if it is a
recent (maintained?) version of gcc, then build stage1 with -O2.

I did some rough measurements on gcc112:

With:
../gcc/configure --disable-libstdcxx-pch --disable-libvtv --disable-libitm
--disable-libcilkrts --disable-libssp --disable-libgomp --disable-werror
--disable-multilib --enable-languages=c,c++,fortran

I get for the current default:
make -j160  28473.74s user 481.93s system 2067% cpu 23:20.53 total

Adding -O2 during stage1 almost halves the bootstrap time:
make -j160  15233.80s user 446.17s system 1918% cpu 13:37.25 total

[Bug bootstrap/63933] Build stage1 with -O2 during bootstrap if host compiler is a recent gcc version

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63933

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |WONTFIX

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
This is on purpose and how do we not know if the original compiler does not
miscompile stage1 at -O2 until it is too late. Closing as won't fix.

[Bug bootstrap/63933] Build stage1 with -O2 during bootstrap if host compiler is a recent gcc version

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63933

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
   Last reconfirmed|                            |2014-11-18
         Resolution|WONTFIX                     |---
     Ever confirmed|0                           |1

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
We dont know, but we will eventually figure it out.  It's a matter of
documentation to try STAGE1_CFLAGS=-g.  Not that we didn't have wrong-code
bugs at -O0 ...

[Bug bootstrap/63933] Build stage1 with -O2 during bootstrap if host compiler is a recent gcc version

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63933

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
Btw, non-GCC host compilers use their default optimization setting which is
usually optimizing.  We're only pessimizing our own GCC here...

[Bug bootstrap/63933] Build stage1 with -O2 during bootstrap if host compiler is a recent gcc version

[andi-gcc at firstfloor dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63933

Andi Kleen <andi-gcc at firstfloor dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andi-gcc at firstfloor dot org

--- Comment #4 from Andi Kleen <andi-gcc at firstfloor dot org> ---
Perhaps using -Og (or -O1) if available?

I actually like to use unoptimized stage1 gcc to debug things with gdb,

The last time I checked the worst offenders were some of the C++ inlines not
getting inlined, and especially the new RTL code very heavily relies on that.
Perhaps just 

#define inline __attribute__((always_inline)) inline

for stage1 would be good enough to fix the worst.