[distcc] distcc and ccache

Gmail kzjeef at gmail.com
Tue Aug 20 02:52:13 MDT 2013


Don't know there is a better way or not, here is my understanding.

GNU make don't know what cc command it called, either disaccharide or ccache, so if you're calling a big -j,  it will create such number job according you make rules.
If you're using distcc and ccache in same time, there is no better way to make live well together. 
So you either only using distcc with a big -j , or use a small -j with ccache.

Don't sure which is faster in your build env, make be you can do some compare in time. 

- Jiejing -

> 在 2013年8月20日,15:55,tal zilcer <tal.zilcer at gmail.com> 写道:
> hi,
> i want to use distcc and ccache in my cluster(10 servers, 8 cpus in each one) in order to improve build times.
> my question is this:
> because i have about 80 cpus using make -j with value of 80-160 seems right according to information on the web.
> but when using ccache if i get only cache hits, distcc wont run(im using CCACHE_PREFIX=distcc) and i will have 80 processes of ccache running on my local host.
> this is much more than the cpus can handle.
> should i use -j as the numbers of cpus on my local host?
> i want to take advantage of all the cpus in the cluster.....
> thanks,
> tal
> __
> 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/20130820/ea438f2f/attachment.html>

More information about the distcc mailing list