Em Fri, Jun 28, 2019 at 10:02:55AM +0200, Jiri Olsa escreveu: > On Fri, Jun 28, 2019 at 05:22:57PM +0800, Jin Yao wrote: > > In some cases small changes in hot loops can show big differences. > > But it's difficult to identify these differences. > > > > perf diff currently can only diff symbols (functions). We can also expand > > it to diff cycles of individual programs blocks as reported by timed LBR. > > This would allow to identify changes in specific code accurately. > > > > With this patch set, for example, > > > > $ perf record -b ./div > > $ perf record -b ./div > > $ perf diff -c cycles > > > > # Event 'cycles' > > # > > # Baseline [Program Block Range] > > Cycles Diff Shared Object Symbol > > # ........ > > ...................................................................... > > ................ .................................. > > # > > 48.75% [div.c:42 -> > > div.c:45] 147 div [.] main > > 48.75% [div.c:31 -> > > div.c:40] 4 div [.] main > > 48.75% [div.c:40 -> > > div.c:40] 0 div [.] main > > 48.75% [div.c:42 -> > > div.c:42] 0 div [.] main > > 48.75% [div.c:42 -> > > div.c:44] 0 div [.] main > > 19.02% [random_r.c:357 -> > > random_r.c:360] 0 libc-2.23.so [.] __random_r > Reviewed-by: Jiri Olsa <jo...@kernel.org>
Looks really nice, thanks, applied. - Arnaldo
