[distcc] Distcc pump mode

[yudhistir reddy]

I have a query ,

Client machine(where build is running) : 16 CPU dual socket ( logically 32
Cores )

Server1 (distccd running ) : 16 CPU dual socket ( logically 32 Cores )

Server2 (distccd running ): 16 CPU dual socket ( logically 32 Cores )

How much -j value i can give so that it would give best performance .
for running single build on client ?
for running two concurrent builds on client ?


might be silly question , but that would be help me reducing my testing
time to give demo .


Regards
Yudhistir






On Fri, Mar 14, 2014 at 4:55 PM, yudhistir reddy
<yudhistir.reddy at gmail.com>wrote:

> Thank you so much for helping me . This issue is resolved . Now i am
> testing the pump mode .
>
> Regards
> Yudhsitir
>
>
> On Fri, Mar 14, 2014 at 12:07 AM, Fergus Henderson <fergus at google.com>wrote:
>
>>
>> On 13 Mar 2014 13:29, "yudhistir reddy" <yudhistir.reddy at gmail.com>
>> wrote:
>> >
>> > # May need to override this (eg. for logging)
>> > exec_cmd() {
>> >     exec "$@"
>> > }
>> >
>> > # This is my official name
>> > MYNAME=mk
>> >
>> > Mydir=`dirname $0`
>> >
>> > # for customization
>> > unset mk_cmd_hooks mk_setup_hooks mk_finish_hooks
>> >
>> > # override this for more complex requirements
>> > mk_target_machine() {
>> >     MACHINE=$1
>> >     export MACHINE
>> > }
>> >
>> > . $Mydir/sb-env.sh
>>
>> What's in sb-env.sh?
>> That looks suspicious...
>>
>> > # for backwards compatability
>> > mk_cmd_hook() { :; }
>> > add_hooks mk_cmd_hooks mk_cmd_hook
>> >
>> > run_hooks mk_setup_hooks
>> > run_hooks mk_cmd_hooks
>> > run_hooks mk_cmd_hooks_lifo LIFO
>> > run_hooks mk_finish_hooks
>> >
>> > # we generally cannot afford to consume the command line
>> > # since virtually any option might be for $cmd
>> > # but we'll make an exception for these
>> > while :
>> > do
>> >     case "$1" in
>> >     --) shift; break;;          # stop trying to interpret anything
>> >     --doc) __doc;;              # does not return
>> >     --help) __help;;            # does not return
>> >     --machine) mk_target_machine $2; shift 2;;
>> >     *) break;;
>> >     esac
>> > done
>> >
>> > case $Myname in
>> > mk|mk-*|*make)
>> >         nice=${nice:-nice}
>> >         cmd=${cmd:-${REAL_MAKE:-/usr/bin/make}};;
>> > *)      cmd=${cmd:-$Myname.real};;
>> > esac
>> > # at last! do some work
>> > exec_cmd $nice $cmd "$@"
>> >
>> >
>> > On Thu, Mar 13, 2014 at 5:31 PM, Vanush <misha at cs.nuim.ie> wrote:
>> >>
>> >> On Thu, Mar 13, 2014 at 05:23:14PM +0530, yudhistir reddy wrote:
>> >> > Sorry for that , that was silly mistake by me .
>> >> >
>> >> > Here is the use case , i did :
>> >> >
>> >> > 1. eval `pump --startup`
>> >> > checked pump mode variables , i can see INCLUDE_SERVER_PORT set to
>> >> > /tmp/distcc-pump.LIlIcs/socket
>> >> >
>> >> >
>> >> > 2. started build
>> >> > mk kernel -j30 WITH_DISTCC=yes
>> >> > Build finished with pump mode  ( But actually it is turing back to
>> Normal
>> >> > mode ) .
>> >> > I see the warning messages in distcclog as like below
>> >>
>> >> can you show us your "mk" wrapper?
>> >>
>> >> Misha
>> >>
>> >>
>> >> --
>> >> Vanush "Misha" Paturyan
>> >> Senior Technical Officer
>> >> Room 1.37
>> >> Computer Science Department
>> >> NUI Maynooth
>> >
>> >
>> >
>> > __
>> > distcc mailing list            http://distcc.samba.org/
>> > To unsubscribe or change options:
>> > https://lists.samba.org/mailman/listinfo/distcc
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.samba.org/pipermail/distcc/attachments/20140314/5501dc98/attachment-0001.html>