From: Junio C Hamano Date: Sun, 18 Mar 2007 09:45:12 GMT Subject: Re: [PATCH 3/2] Avoid unnecessary strlen() calls Message-ID: <7v8xdunavr.fsf@assigned-by-dhcp.cox.net> In-Reply-To: Linus Torvalds writes: > This shaves off another ~5-10% from some loads that are very tree > intensive (notably doing commit filtering by a pathspec). > > Signed-off-by: Linus Torvalds " With your 256-entry cache, Nico's reusing objects out of delta base cache, and this strlen() patch git blame -C block/ll_rw_blk.c gets these numbers: (v1.5.0) 14.71user 0.26system 0:15.07elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+93622minor)pagefaults 0swaps (master + three patches) 8.94user 0.14system 0:09.10elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+40075minor)pagefaults 0swaps Just for fun, these are the same for the kernel history with tglx-history repository's history grafted behind it, i.e. with this grafts file: $ cat .git/info/grafts 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 e7e173af42dbf37b1d946f9ee00219cb3b2bea6a (v1.5.0) 73.80user 2.57system 1:16.40elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+773077minor)pagefaults 0swaps (master + three patches) 65.14user 0.40system 1:05.55elapsed 99%CPU (0avgtext+0avgdata 0maxresident)k 0inputs+0outputs (0major+125052minor)pagefaults 0swaps In either case, it is showing drastic reduction of minor faults.