Subject: Re: i386 MP performance (generic.mp and generic almost same)
To: None <current-users@netbsd.org>
From: Cem Kayali <cemkayali@eticaret.com.tr>
List: current-users
Date: 11/11/2007 07:52:38
Hi,
using "-j 4"
build.sh command: ./build.sh -j 4 -O ../obj -T ../tools -U -u
kernel=3DGENERIC
build.sh started: Sun Nov 11 17:31:30 EET 2007
build.sh ended: Sun Nov 11 17:36:21 EET 2007
4 mins 51 secs. Max server load was ~5.00
without "-j 4"
build.sh command: ./build.sh -O ../obj -T ../tools -U -u
kernel=3DGENERIC
build.sh started: Sun Nov 11 17:39:43 EET 2007
build.sh ended: Sun Nov 11 17:47:37 EET 2007
7 mins 54 secs. Max server load was ~1.50
Ratio =3D 291 / 474 =3D 0,61
--------------------------------------------
I'm sorry to hear i can use only single core especially for compiling pkgsr=
c
softwares...=20
Regards,
Paul Goyette wrote:
>=20
> On Sun, 11 Nov 2007, Steven M. Bellovin wrote:
>=20
>> Compilation is indeed CPU-bound; however, unless you've taken special
>> steps, it's a single process at a time. What you need to do is specify
>> '-j 2' when invoking build.sh, to run two processes simultaneously;
>> see /usr/src/BUILDING for details.
>=20
> In my experience, '-j 2' won't be enough to provide definitive results,=
=20
> since that seems to improve performance on uni-processor systems, too=20
> (though not by as much). On my AMD X2 systems I run builds with it set=
=20
> to '-j 4' and that seems to give the most benefit.
>=20
> But even on uni-processors '-j 2' will improve things since it gives a=20
> chance for a second job to run on the (only) CPU while the other one is=
=20
> waiting for I/O completion.
>=20
> ----------------------------------------------------------------------
> | Paul Goyette | PGP DSS Key fingerprint: | E-mail addresses: |
> | Customer Service | FA29 0E3B 35AF E8AE 6651 | paul@whooppee.com |
> | Network Engineer | 0786 F758 55DE 53BA 7731 | pgoyette@juniper.net |
> ----------------------------------------------------------------------
>=20
>=20
-----
=20
=20
=20
----------------
Cem Kayali / Kayal=C4=B1
=20
--=20
View this message in context: http://www.nabble.com/i386-MP-performance-%28=
generic.mp-and-generic-almost-same%29-tf4784539.html#a13692625
Sent from the current-users mailing list archive at Nabble.com.