Sorted by build-time
- openoffice.org - 7h02
- gcc-4.2 - 5h58
- gcc-snapshot - 4h17
- gcc-4.1 - 3h30
- ghc-cvs - 3h16
- linux-2.6 - 2h34
- gpc-4.1 - 2h25
- gclcvs - 2h21
- glibc - 2h15
- python2.5 - 2h09
- axiom - 2h07
- gnat-4.1 - 2h06
- gcj-4.1 - 1h52
- gromacs - 1h51
- mysql-dfsg-5.0 - 1h46
- acl2 - 1h44
- gcj-4.2 - 1h34
- ghc6 - 1h33
- python-kde3 - 1h32
- aegis - 1h25
Sorted by build-space
- openoffice.org - 7217MB
- nel - 2773MB
- gcc-snapshot - 2499MB
- koffice - 2247MB
- wxwidgets2.6 - 2151MB
- linux-2.6 - 2150MB
- webkit - 2059MB
- glibc - 2028MB
- xulrunner - 2023MB
- iceape - 2007MB
- boost - 1983MB
- mesa - 1943MB
- gcc-4.2 - 1943MB
- qt4-x11 - 1915MB
- python-kde3 - 1847MB
- gcc-4.1 - 1828MB
- gcj-4.1 - 1783MB
- eclipse - 1763MB
- iceweasel - 1727MB
- gcj-4.2 - 1722MB
No real surprises here. The data was collected from the Buildd logs using the latest "maybe-successful" i386 or amd64 build. Factoring into that the varied nature of the buildd machine specifications make this is data somewhat unreliable.
It would be nice if someone could obtain (and post) results from a more, umm, controlled environment.