[Xorp-hackers] Size of xorps binaries
Mark Handley
Mark Handley <M.Handley@cs.ucl.ac.uk>
Sat, 10 Sep 2005 14:56:32 +0100
------=_Part_3266_2355960.1126360592582
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
I think for developers, it's something we probably have to accept. For=20
operational use, the first thing to do is build with the optimiser (to=20
improve speed not size), and strip the binaries (to improve size). This=20
makes a huge difference:
echidna.cs.ucl.ac.uk <http://echidna.cs.ucl.ac.uk>: ls -l xorp_bgp=20
-rwxr-xr-x 1 mjh mjh 34545650 Jul 13 18:10 xorp_bgp
echidna.cs.ucl.ac.uk <http://echidna.cs.ucl.ac.uk>: strip xorp_bgp=20
echidna.cs.ucl.ac.uk <http://echidna.cs.ucl.ac.uk>: ls -l xorp_bgp
-rwxr-xr-x 1 mjh mjh 2840504 Sep 10 14:55 xorp_bgp
We've discussed off-and-on over the years using shared libraries. So far,=
=20
we've not taken the plunge - it makes some things more complex, but it woul=
d=20
probably be essential if you wanted a relatively small memory footprint.=20
More non-embedded uses though, stripping the binaries seems to be=20
sufficient.
- Mark
On 9/10/05, Kristian Larsson <kristian@juniks.net> wrote:
>=20
> I just noticed after compiling xorp on my Linux
> laptop the size of the binaries are approximately
> twice that of the binaries compiled on a FreeBSD
> machine. Why is this?
> The xorpsh is just huge - 26MB! And the bgp
> process weighs in at a hefty 44MB.
>=20
> Could this also be the reason it takes ~30 seconds
> to start the fea, rib and an ospf process?
> IMHO that's a very long time.
>=20
> Anything that could be fixed or is it something
> you just have to accept?
>=20
> //Kristian
>=20
> _______________________________________________
> Xorp-hackers mailing list
> Xorp-hackers@icir.org
> http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-hackers
>
------=_Part_3266_2355960.1126360592582
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
I think for developers, it's something we probably have to
accept. For operational use, the first thing to do is build with
the optimiser (to improve speed not size), and strip the binaries (to
improve size). This makes a huge difference:<br>
<br>
<a href=3D"http://echidna.cs.ucl.ac.uk">echidna.cs.ucl.ac.uk</a>: ls -l xor=
p_bgp <br>
-rwxr-xr-x 1 mjh mjh 34545650 Jul 13 18:10 xorp_bgp<br>
<a href=3D"http://echidna.cs.ucl.ac.uk">echidna.cs.ucl.ac.uk</a>: strip xor=
p_bgp <br>
<a href=3D"http://echidna.cs.ucl.ac.uk">echidna.cs.ucl.ac.uk</a>: ls -l xor=
p_bgp<br>
-rwxr-xr-x 1 mjh mjh 2840504 Sep 10 14:55 xorp_bgp<br>
<br>
<br>
We've discussed off-and-on over the years using shared libraries.
So far, we've not taken the plunge - it makes some things more complex,
but it would probably be essential if you wanted a relatively small
memory footprint. More non-embedded uses though, stripping the
binaries seems to be sufficient.<br>
<br>
- Mark<br><br><div><span class=3D"gmail_quote">On 9/10/05, <b class=
=3D"gmail_sendername">Kristian Larsson</b> <<a href=3D"mailto:kristian@j=
uniks.net">kristian@juniks.net</a>> wrote:</span><blockquote class=3D"gm=
ail_quote" style=3D"border-left: 1px solid rgb(204, 204, 204); margin: 0pt =
0pt 0pt 0.8ex; padding-left: 1ex;">
I just noticed after compiling xorp on my Linux<br>laptop the size of the b=
inaries are approximately<br>twice that of the binaries compiled on a FreeB=
SD<br>machine. Why is this?<br>The xorpsh is just huge - 26MB! And the bgp
<br>process weighs in at a hefty 44MB.<br><br>Could this also be the reason=
it takes ~30 seconds<br>to start the fea, rib and an ospf process?<br>IMHO=
that's a very long time.<br><br>Anything that could be fixed or is it some=
thing
<br>you just have to accept?<br><br>//Kristian<br><br>_____________________=
__________________________<br>Xorp-hackers mailing list<br><a href=3D"mailt=
o:Xorp-hackers@icir.org">Xorp-hackers@icir.org</a><br><a href=3D"http://mai=
lman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-hackers">
http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-hackers</a><br></blo=
ckquote></div><br>
------=_Part_3266_2355960.1126360592582--