got it. It's JMH based test bench. We will look into it. On Tue, Apr 24, 2018 at 1:07 PM, whanice <ywh....@gmail.com> wrote:
> I use 100K String . and just call readString. Here is the simple code: > > @Benchmark > public void deserial(BenchmarkState state) throws Exception { > ByteArrayInputStream bin = new ByteArrayInputStream(state.bytes); > Hessian2Input input = new Hessian2Input(bin); > String deserialize = input.readString(); > Assert.assertEquals(state.str, deserialize); > } > > public static void main(String[] args) throws RunnerException { > Options options = new OptionsBuilder() > .include(HessianLitePerfTest.class.getSimpleName()) > .include(Hessian4PerfTest.class.getSimpleName()) > .warmupIterations(10) > .measurementIterations(30) > .threads(10) > .mode(Mode.Throughput) > .timeUnit(TimeUnit.SECONDS) > .forks(1) > .build(); > > new Runner(options).run(); > } > > > > > > > 2018-04-24 11:10 GMT+08:00 Ian Luo <ian....@gmail.com>: > > > Hessian-lite is tailored from the original hessian in order to keep it > > lightweight. From your test, it looks hessian 4 performs much much better > > than hessian-lite and hessian 3. Would you mind to describe how you run > the > > test, or even better, share us your test code? Then we will consider to > > upgrade to hessian 4 and deprecate hessian-lite. > > > > Thanks, > > -Ian. > > > > > > On Mon, Apr 23, 2018 at 8:01 PM, whanice <ywh....@gmail.com> wrote: > > > > > I made a simply performance test in my mac to messure hessian. > > > the result is > > > > > > Benchmark Mode Cnt Score Error Units > > > Hessian4PerfTest.deserial thrpt 30 4068.415 ± 189.925 ops/s > > > HessianLitePerfTest.deserial thrpt 30 1941.054 ± 77.005 ops/s > > > > > > We can see hessian4(4.0.37) performs much better than > hessian-lite(3.2.2) > > > in deserialization. > > > > > > So here comes the question > > > > > > - 1. What is the difference between hessian-lite(3.2.2) and > hessian3? > > > - 2. Can sync with hessian4? > > > > > >