On Thu, Jul 28, 2011 at 03:25:48PM -0500, Paul Hartman wrote:
> > On my machine, it took this:
> >
> > root@fireball / # genlop -t libreoffice
> > * app-office/libreoffice
> >
> > Thu Jul 28 12:33:11 2011 >>> app-office/libreoffice-3.3.1
> > merge time: 1 hour, 9 minutes and 33 seconds.
> >
> > root@fireball / #
> >
> > I did stop it with a ctrl Z for about 5 minutes. I was deleting stuff to
> > give it some more room. This is OOo:
> >
> > Tue Jul 5 05:15:01 2011 >>> app-office/openoffice-3.2.1-r1
> > merge time: 50 minutes and 27 seconds.
> >
> >
> > That's about a average. Some were binary installs. I can't recall why I
> > did that now but it was since it only took a minute or so.
> >
> > That's the report from this rig. AMD 4 cores running at 3.2Ghz with 16Gbs
> > of ram. No tmpfs this time. That wouldn't be fair since I had to stop it
> > for a few minutes.
> >
> > Dale
>
> Here is mine:
>
> Wed May 25 10:07:18 2011 >>> app-office/libreoffice-3.3.2
> merge time: 34 minutes and 10 seconds.
>
> On my Intel Core i7 920 with 12GB of RAM using tmpfs. :)
>
> Do those merge times include download time? I wonder...
>
BTW, I was emerging libreoofice yesterday, and it took ~3.5 hours,
although OOo usually took ~1 hour. It seemed as if it wasn't doing
anything paralell, so I blamed my MAKEOPTS="-j -l4" (to which I changed
some time ago, but all the OOo compiles were before that). I just tried
today again with MAKEOPTS=-j4 and it took only ~1 hour ;).
Guess dmake doesn't know the --load-average stuff ;)
can't give a nice genlop -t output, as the first time i was building it
first with FEATURES=buildpkgonly and then installing with --usepkg (just
after i removed OOo), as portage didn't wan't to resolve the blocker,
and i didn't want to be without it while it compiled ... ;)
# genlop -t libreoffice
* app-office/libreoffice
Fri Jul 29 08:57:10 2011 >>> app-office/libreoffice-3.3.3
merge time: 1 minute and 59 seconds.
Fri Jul 29 13:55:42 2011 >>> app-office/libreoffice-3.3.3
merge time: 1 hour, 7 minutes and 25 seconds.
the first one is for the --usepkg binary merge ;), but I checked the
timestamps in emerge.log and the compile was a small bit over 3.5h
yoyo
