> On June 28, 2018, 2:04 p.m., Benjamin Mahler wrote: > > Looks pretty good! The only thing that feels a little off is the size > > parameterization. It currently represents the number of ranges, but I > > imagine other benchmarks might not want to use that as the "size" (vs the > > overall range size or number of items). I think it's fine for now but > > probably warrants some explanation (e.g. we choose to parameterize on > > number of subranges becacuse it's a dominant factor in the performance of > > operations). What do you think?
Sounds good. Comment added. > On June 28, 2018, 2:04 p.m., Benjamin Mahler wrote: > > src/tests/resources_tests.cpp > > Lines 3981 (patched) > > <https://reviews.apache.org/r/67590/diff/2/?file=2046953#file2046953line4000> > > > > Oh, I guess I misunderstood this value, it's not the number of items > > but rather the number of sub-ranges? Seems like this needs to at least be > > documented? I guess this means 10,000 represents far more than 10,000 ports > > in the overlapping benchmark? Probably warrants writing it down? Correct. Added comment. - Meng ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/67590/#review205529 ----------------------------------------------------------- On June 28, 2018, 5:38 p.m., Meng Zhu wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/67590/ > ----------------------------------------------------------- > > (Updated June 28, 2018, 5:38 p.m.) > > > Review request for mesos and Benjamin Mahler. > > > Bugs: MESOS-8989 > https://issues.apache.org/jira/browse/MESOS-8989 > > > Repository: mesos > > > Description > ------- > > See summary. > > Also removed the current range benchmark which could produce > misleading results for subtraction (See MESOS-8989). > > > Diffs > ----- > > src/tests/resources_tests.cpp 0095b945ebfd9be52162547f3999826c8cc33f87 > > > Diff: https://reviews.apache.org/r/67590/diff/4/ > > > Testing > ------- > > Build with `-O2` optimization, ran on a multicore machine with peak frequency > at 2.2GHz: > > Took 1.504607ms to perform 1000 'a += b' operations on ports:[1-6, 11-16, > 21-26...91-96] and ports:[3-8, 13-18, 23-28..., 93-98] with 10 sub-ranges > Took 7.119324ms to perform 1000 'a -= b' operations on ports:[1-6, 11-16, > 21-26...91-96] and ports:[3-8, 13-18, 23-28..., 93-98] with 10 sub-ranges > Took 2.991275ms to perform 1000 'a + b' operations on ports:[1-6, 11-16, > 21-26...91-96] and ports:[3-8, 13-18, 23-28..., 93-98] with 10 sub-ranges > Took 8.681909ms to perform 1000 'a - b' operations on ports:[1-6, 11-16, > 21-26...91-96] and ports:[3-8, 13-18, 23-28..., 93-98] with 10 sub-ranges > > Took 14.814339ms to perform 1000 'a += b' operations on ports:[1-6, 11-16, > 21-26...991-996] and ports:[3-8, 13-18, 23-28..., 993-998] with 100 sub-ranges > Took 80.680714ms to perform 1000 'a -= b' operations on ports:[1-6, 11-16, > 21-26...991-996] and ports:[3-8, 13-18, 23-28..., 993-998] with 100 sub-ranges > Took 25.640379ms to perform 1000 'a + b' operations on ports:[1-6, 11-16, > 21-26...991-996] and ports:[3-8, 13-18, 23-28..., 993-998] with 100 sub-ranges > Took 89.253517ms to perform 1000 'a - b' operations on ports:[1-6, 11-16, > 21-26...991-996] and ports:[3-8, 13-18, 23-28..., 993-998] with 100 sub-ranges > > Took 234.144019ms to perform 1000 'a += b' operations on ports:[1-6, 11-16, > 21-26...9991-9996] and ports:[3-8, 13-18, 23-28..., 9993-9998] with 1000 > sub-ranges > Took 892.338197ms to perform 1000 'a -= b' operations on ports:[1-6, 11-16, > 21-26...9991-9996] and ports:[3-8, 13-18, 23-28..., 9993-9998] with 1000 > sub-ranges > Took 317.301614ms to perform 1000 'a + b' operations on ports:[1-6, 11-16, > 21-26...9991-9996] and ports:[3-8, 13-18, 23-28..., 9993-9998] with 1000 > sub-ranges > Took 1.024019112secs to perform 1000 'a - b' operations on ports:[1-6, 11-16, > 21-26...9991-9996] and ports:[3-8, 13-18, 23-28..., 9993-9998] with 1000 > sub-ranges > > > Thanks, > > Meng Zhu > >
