-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am 07.01.2015 um 21:17 schrieb Neil Bothwick:
> That's not what I meant, but I see your point about using --reflink > if making copies. My thought was to forget the whole tmpfs and > copying think, set KEEP WORK in FEATURES and use XFS for > PORTAGE_TMPDIR. That way most of the work is taking place in the > cache without the frequent disk writes of other filesystems. I > would expect XFS to be faster for this job, but have no data to > support that assumption. Maybe I am not fully up to date with this thread. You mean, PORTAGE_TMPDIR on xfs (on disk ... hdd or ssd) might be faster than on tmpfs? Interesting ... I run that on tmpfs for a while now, with 8GB of RAM dedicated ... I never benchmarked, but always was confident that this was the fastest possible setup. Or is that statement pointed at the specific case of building and keeping binary packages at the end of emerging (which I assume) ? Stefan -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIcBAEBAgAGBQJUrZjaAAoJEClcuD1V0PzmTd0P/jg+1jaOzQZ63HVPBg75bzfh jpH0u+SIQ7G8ceNI2fXGQc5QdwnNPdU0bRGd0Sdg+jFc+zGHaHBFVcWXgxzuxPWD IVIQ73TH9lxYIw4cjD3iLwcgeFCNBIMyhok/1UOzYq6qDO2HDx4yl2W7BSp4+pw8 AIlK9NwJ3ysMPC97mHdnAKr7DN2JqXw5bEDA6ssiyKtYh+KnNl46IYadtdbDk1wz C8KODjpXYrtve9h1iLvpZicI+unnBWu10B3sEp7I3EoJz2xGGys82hlxXQJ6f1v4 kWR1Ej7P5ttjOpefDcyLVi7I4MEsZAS6zHVmkqUkdm1eq/bE28L6yeszC4+ypSrh l1dYvYfIMOz/81Bu7bqLei/FbBFSe2PnZg7YJnBb3WJweiyi4TFnPxYo1tmhUaZa NhFj8z8YcmLS9XzVHSqb5tmgc0HHmid7xT3ws39fOpJaJzcP3YnVodQVLswAvrdV pMIPFzwwiPvmHxnXHtbxnV/Z+H0Ka3qRrLkIVwKSNQVQlWa5Z5kRmSOovNAmjfLh BVITUZH5V6dY1fr6GwtvXuz7fwgaS9us5FhrSgJ5bTFXI/+sDyHl9oKfUZqOomt4 8XjTtcc6iHIV9+3MLJJfc21AiMFi7T00TUQcOrdrYMWnr+AhJOSIQaZCBJbxxp7O GOXATsnat7oPgbAMsqi0 =WzQB -----END PGP SIGNATURE-----
