Review Rules and staticly linked packages agains dietlibc
Hans de Goede
j.w.r.degoede at hhs.nl
Fri Feb 24 18:05:59 UTC 2006
Enrico Scholz wrote:
> ok, completing the numbers and assuming a dietlibc less world, we will
> have to compile 'r' dynamically which results into the following
> datasets:
>
> user-time:
> res.ds.* res.diet.* res.glibc.* resD.glibc.*
> 0m0.000s(#) 0m0.260s 0m0.356s 0m1.088s
> 0m0.180s 0m0.308s 0m0.344s 0m1.124s
> 0m0.220s 0m0.300s 0m0.348s 0m1.140s
> 0m0.160s 0m0.328s 0m0.364s 0m1.172s
> 0m0.212s 0m0.288s 0m0.404s 0m1.060s
> 0m0.208s 0m0.296s 0m0.332s 0m1.072s
> 0m0.148s 0m0.348s 0m0.280s 0m1.108s
> 0m0.172s 0m0.308s 0m0.304s 0m1.080s
> 0m0.224s 0m0.324s 0m0.292s 0m0.972s
> 0m0.212s 0m0.312s 0m0.368s 0m1.092s
> 0m0.212s 0m0.280s 0m0.352s 0m1.000s
> ------------------------------------------------------
> 0m0.195s 0m0.305s 0m0.340s 0m1.083s
> 100% 156% 175% 555%
>
>
> sys-time
> res.ds.* res.diet.* res.glibc.* resD.glibc.*
> 0m0.000s(#) 0m1.516s 0m2.248s 0m4.100s
> 0m1.212s 0m1.748s 0m2.124s 0m4.084s
> 0m1.512s 0m1.412s 0m2.160s 0m3.988s
> 0m1.296s 0m1.672s 0m2.204s 0m4.060s
> 0m1.240s 0m1.764s 0m2.192s 0m3.444s
> 0m1.340s 0m1.748s 0m2.132s 0m3.804s
> 0m1.444s 0m1.716s 0m1.944s 0m4.040s
> 0m1.612s 0m1.720s 0m1.980s 0m4.044s
> 0m1.428s 0m1.808s 0m1.924s 0m3.688s
> 0m1.492s 0m1.804s 0m1.968s 0m3.864s
> 0m1.480s 0m1.640s 0m2.120s 0m4.152s
> ------------------------------------------------------
> 0m1.406s 0m1.686s 0m2.090s 0m3.933s
> 100% 119% 149% 279%
>
>
Yes, all very interesting but you're using a static inetd like in.xxx
program for these tests, unless you want to argue that all programs
executed by invsd are going to be static linked too, in which case you
can start to vastly expand you're auditing of dietlibc to all
functionality such a in.xxx program could use, so basicly to all
functionality I would like to see results for resD.diet and resD.ds and
compare those to resD.glibc
About your resident sizes results this doesn't take the shared part of
this into the picture. Please substract the shared part (minus the
stripped binary size) from these figures since the few basic glibc
functions used will be resident already in any realistic Fedora test
scenario.
Regards,
Hans
More information about the fedora-extras-list
mailing list