https://bugzilla.redhat.com/show_bug.cgi?id=1361659 Nir Soffer <nsoffer@xxxxxxxxxx> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags| |needinfo?(dougsland@redhat. | |com) --- Comment #21 from Nir Soffer <nsoffer@xxxxxxxxxx> --- (In reply to Douglas Schilling Landgraf from comment #7) > #2 Use %{?_smp_mflags} > make %{?_smp_mflags} DESTDIR=%{buildroot} install I tested this on my laptop: $ for i in `seq 10`; do rm -rf /home/nsoffer/tmp/root/*; /usr/bin/time -f %E make -j4 DESTDIR=/home/nsoffer/tmp/root install 2>&1 | tail -n 1; done 0:04.02 0:04.05 0:04.05 0:04.11 0:03.93 0:03.94 0:03.91 0:04.05 0:03.97 0:04.01 $ for i in `seq 10`; do rm -rf /home/nsoffer/tmp/root/*; /usr/bin/time -f %E make DESTDIR=/home/nsoffer/tmp/root install 2>&1 | tail -n 1; done 0:04.19 0:04.31 0:04.21 0:04.48 0:04.34 0:04.20 0:04.27 0:04.49 0:04.26 0:04.40 I don't see any significant improvement, but I see risk of races during install, so I don' think this is a good idea. $ for i in `seq 10`; do make clean >/dev/null 2>&1; /usr/bin/time -f %E make -j4 2>&1 | tail -n 1; done 0:01.06 0:01.04 0:01.08 0:01.22 0:01.07 0:01.06 0:01.10 0:01.13 0:01.08 0:01.21 $ for i in `seq 10`; do make clean >/dev/null 2>&1; /usr/bin/time -f %E make 2>&1 | tail -n 1; done 0:01.50 0:01.99 0:01.95 0:01.87 0:01.84 0:01.88 0:01.75 0:02.11 0:01.77 0:01.71 This is significantly faster, but it save less then a second, so I don't think we should use it. This also breaks when updating the spec, so it requires more debugging if we do want this sub-second improvement. -- You are receiving this mail because: You are on the CC list for the bug. You are always notified about changes to this product and component _______________________________________________ package-review mailing list package-review@xxxxxxxxxxxxxxxxxxxxxxx https://lists.fedoraproject.org/admin/lists/package-review@xxxxxxxxxxxxxxxxxxxxxxx