We discovered that the first benchmarks were run with DISTCC_DIR defaulting to an NFS directory, which killed the perfomance. Here are some more real numbers for a linux kernel build, courtesy of Dongmin Zhang:
155 sec test system without distcc. 44 sec. normal distcc, eight-server cluster, -j8 46 sec. caching distcc on same cluster, first run 35 sec caching distcc on same cluster, second run So with a perfectly preloaded cache, distcc went from 3x faster to 4x faster. Switching to a faster hash code (adler32 instead of sha-1) didn't change the results. We're still playing around with this here, but it might be ready for other folks to try and/or review. Anyone interested? - Dan -- Why won't Johnny run Linux? See http://kegel.com/linux/comfort __ distcc mailing list http://distcc.samba.org/ To unsubscribe or change options: https://lists.samba.org/mailman/listinfo/distcc