There is a tool out there that I wrote a long time ago called DMUCS, which is a wrapper around distcc. I built it for just this situation -- where you have some machines that are much faster than others in your compilation farm.
dmucs.sourceforge.net is where you can find it. I have not maintained this tool, I'm sorry to say. I've spent some time rewriting it in python, but haven't finished that... What DMUCS does is run a little daemon on each compilation machine, monitoring that machine's load average. Also, you configure DMUCS to know that some machines are x amount faster than other machines. Then, in your makefile, you run the dmucs tool, which find the best machine at the time -- i.e., the fastest machine, unless that fastest machine's load average is already high -- and tells distcc to use the best machine for that compilation. I think dmucs still does work -- but you might have a bit of a time compiling it, as it is getting pretty old now. let me know if you are more interested, and/or need help. On Wed, Jul 18, 2012 at 3:07 PM, Αναστασίου Γεώργιος < [email protected]> wrote: > Yes they are, i can see that with distccmon,but i am not sure about the > percentage of use each pc uses for compiling,even if the order is from > stronger to weaker machine. > > > On 18 July 2012 22:00, Martin Pool <[email protected]> wrote: > >> Have you looked at the logs to check the jobs are actually being >> distributed? >> > > > __ > distcc mailing list http://distcc.samba.org/ > To unsubscribe or change options: > https://lists.samba.org/mailman/listinfo/distcc > -- Prof. Victor Norman Computer Science Dept., Calvin College [email protected] / 616 526-7805 “A designer knows he has achieved perfection not when there is nothing left to add, but when there is nothing left to take away.” -- Antoine de Saint Exupéry
__ distcc mailing list http://distcc.samba.org/ To unsubscribe or change options: https://lists.samba.org/mailman/listinfo/distcc
