On 15 July 2010 16:38, Robert Miller <r...@rlmiller.org> wrote: > On Thu, Jul 15, 2010 at 1:51 PM, David Kirkby <david.kir...@onetel.net> wrote: >> 3) GLPK should be built before R if we ever hope to use it in R >> 4) GLPK should be built before CVXOPT if there is any hope of using >> those together. > > Shouldn't we wait until we implement using GLPK from these? Otherwise > we are potentially slowing down parallel builds... > > > > -- > Robert L. Miller > http://www.rlmiller.org/
Theoretically that is true. But GLPK builds really quickly - in under 6 seconds on my computer, using a little over 15 s of CPU time. drkir...@hawk:~/sage-4.5.rc0$ ./sage -f glpk-4.44 <SNIP> real 0m5.295s user 0m11.287s sys 0m4.012s Successfully installed glpk-4.44 That's on a 3.33 GHz Sun Ultra 27, but it is still a very small fraction of the time it takes to build Sage, which is around 30 minutes, though R and Maxima are not building, so I rekon it would be closer to 35 or 40 minutes. That makes the time to build GLPK less about 0.25% of the total time to build Sage. My logic for suggesting doing them all at once was: * Updating 'deps' is a bit of a pain, since its not under revision control, so it would be worth getting all the changes in at one time. * It would make testing the integration a bit easier, if the dependencies were already in place. Perhaps I'm missing something, and the penalty to build Sage could be higher than the time it takes to build GLPK, but I don't think that would be the case. In fact, I suspect it would be less than that. Dave -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
