| Message ID | alpine.DEB.2.20.13.1607201009280.1419@idea |
|---|---|
| State | New |
| Headers | show |
On Wed, Jul 20, 2016 at 7:11 AM, Patrick Palka <patrick@parcs.ath.cx> wrote: > On Wed, 20 Jul 2016, Bernd Schmidt wrote: > >> On 07/19/2016 10:20 AM, Richard Biener wrote: >> > I like it. Improving re-build time in my dev tree is very much >> > welcome, and yes, >> > libbackend build time is a big part of it usually (plus of course cc1 >> > link time). >> >> Since that wasn't an entirely explicit ack, I'll add mine. Thank you for doing >> this. >> >> >> Bernd >> >> > > Committed as r238524 with the following minor change to the configure > test to use $CFLAGS and $LDFLAGS consistently: What is interesting is I did not see any change in my bootstrap/test times. Constant at 1 hour and 39 minutes (this includes archiving the build away). And yes I checked the logs to make sure T option to ar is being used too. Thanks, Andrew > > diff --git a/gcc/configure.ac b/gcc/configure.ac > index 63052ba..241e82d 100644 > --- a/gcc/configure.ac > +++ b/gcc/configure.ac > @@ -4905,7 +4905,7 @@ echo 'int main (void) { return 0; }' > conftest.c > if ($AR --version | sed 1q | grep "GNU ar" \ > && $CC $CFLAGS -c conftest.c \ > && $AR rcT conftest.a conftest.o \ > - && $CC -o conftest conftest.a) >/dev/null 2>&1; then > + && $CC $CFLAGS $LDFLAGS -o conftest conftest.a) >/dev/null 2>&1; then > thin_archive_support=yes > fi > rm -f conftest.c conftest.o conftest.a conftest
On Fri, Jul 22, 2016 at 2:19 AM, Andrew Pinski <pinskia@gmail.com> wrote: > On Wed, Jul 20, 2016 at 7:11 AM, Patrick Palka <patrick@parcs.ath.cx> wrote: >> On Wed, 20 Jul 2016, Bernd Schmidt wrote: >> >>> On 07/19/2016 10:20 AM, Richard Biener wrote: >>> > I like it. Improving re-build time in my dev tree is very much >>> > welcome, and yes, >>> > libbackend build time is a big part of it usually (plus of course cc1 >>> > link time). >>> >>> Since that wasn't an entirely explicit ack, I'll add mine. Thank you for doing >>> this. >>> >>> >>> Bernd >>> >>> >> >> Committed as r238524 with the following minor change to the configure >> test to use $CFLAGS and $LDFLAGS consistently: > > What is interesting is I did not see any change in my bootstrap/test > times. Constant at 1 hour and 39 minutes (this includes archiving the > build away). > And yes I checked the logs to make sure T option to ar is being used too. > > Thanks, > Andrew Do you see any change in rebuild times in a --disable-bootstrap dev tree? On average I guess you'd save about 20s per stage which is a significant amount of time during development (in a --disable-bootstrap tree) but not so much when bootstrap/regtesting. The ar+ranlib invocation is primarily I/O bottlenecked (it writes out 900MB of stuff) so the amount of time saved by using a thin archive depends on how slow your block device is. > >> >> diff --git a/gcc/configure.ac b/gcc/configure.ac >> index 63052ba..241e82d 100644 >> --- a/gcc/configure.ac >> +++ b/gcc/configure.ac >> @@ -4905,7 +4905,7 @@ echo 'int main (void) { return 0; }' > conftest.c >> if ($AR --version | sed 1q | grep "GNU ar" \ >> && $CC $CFLAGS -c conftest.c \ >> && $AR rcT conftest.a conftest.o \ >> - && $CC -o conftest conftest.a) >/dev/null 2>&1; then >> + && $CC $CFLAGS $LDFLAGS -o conftest conftest.a) >/dev/null 2>&1; then >> thin_archive_support=yes >> fi >> rm -f conftest.c conftest.o conftest.a conftest
On Fri, Jul 22, 2016 at 8:02 AM, Patrick Palka <patrick@parcs.ath.cx> wrote: > On Fri, Jul 22, 2016 at 2:19 AM, Andrew Pinski <pinskia@gmail.com> wrote: >> On Wed, Jul 20, 2016 at 7:11 AM, Patrick Palka <patrick@parcs.ath.cx> wrote: >>> On Wed, 20 Jul 2016, Bernd Schmidt wrote: >>> >>>> On 07/19/2016 10:20 AM, Richard Biener wrote: >>>> > I like it. Improving re-build time in my dev tree is very much >>>> > welcome, and yes, >>>> > libbackend build time is a big part of it usually (plus of course cc1 >>>> > link time). >>>> >>>> Since that wasn't an entirely explicit ack, I'll add mine. Thank you for doing >>>> this. >>>> >>>> >>>> Bernd >>>> >>>> >>> >>> Committed as r238524 with the following minor change to the configure >>> test to use $CFLAGS and $LDFLAGS consistently: >> >> What is interesting is I did not see any change in my bootstrap/test >> times. Constant at 1 hour and 39 minutes (this includes archiving the >> build away). >> And yes I checked the logs to make sure T option to ar is being used too. >> >> Thanks, >> Andrew > > Do you see any change in rebuild times in a --disable-bootstrap dev > tree? On average I guess you'd save about 20s per stage which is a > significant amount of time during development (in a > --disable-bootstrap tree) but not so much when bootstrap/regtesting. > The ar+ranlib invocation is primarily I/O bottlenecked (it writes out > 900MB of stuff) so the amount of time saved by using a thin archive > depends on how slow your block device is. ... so it's weird that your build times are constant because building a thin archive is much faster than building a regular archive and using a thin archive should not make subsequent steps, like linking that archive, slower. So overall build times should be lower. > >> >>> >>> diff --git a/gcc/configure.ac b/gcc/configure.ac >>> index 63052ba..241e82d 100644 >>> --- a/gcc/configure.ac >>> +++ b/gcc/configure.ac >>> @@ -4905,7 +4905,7 @@ echo 'int main (void) { return 0; }' > conftest.c >>> if ($AR --version | sed 1q | grep "GNU ar" \ >>> && $CC $CFLAGS -c conftest.c \ >>> && $AR rcT conftest.a conftest.o \ >>> - && $CC -o conftest conftest.a) >/dev/null 2>&1; then >>> + && $CC $CFLAGS $LDFLAGS -o conftest conftest.a) >/dev/null 2>&1; then >>> thin_archive_support=yes >>> fi >>> rm -f conftest.c conftest.o conftest.a conftest
Patrick Palka <patrick@parcs.ath.cx> writes: > ... so it's weird that your build times are constant because building > a thin archive is much faster than building a regular archive and > using a thin archive should not make subsequent steps, like linking > that archive, slower. So overall build times should be lower. Linking a thin archive requires opening each archive member as a file instead of only reading within a single file. This could add some non-trivial overhead. Andreas.
diff --git a/gcc/configure.ac b/gcc/configure.ac index 63052ba..241e82d 100644 --- a/gcc/configure.ac +++ b/gcc/configure.ac @@ -4905,7 +4905,7 @@ echo 'int main (void) { return 0; }' > conftest.c if ($AR --version | sed 1q | grep "GNU ar" \ && $CC $CFLAGS -c conftest.c \ && $AR rcT conftest.a conftest.o \ - && $CC -o conftest conftest.a) >/dev/null 2>&1; then + && $CC $CFLAGS $LDFLAGS -o conftest conftest.a) >/dev/null 2>&1; then thin_archive_support=yes fi rm -f conftest.c conftest.o conftest.a conftest