Another important thing to consider is that nano time costs a lot more than System.currentTimeMillis().
I've done some benchmarking in the past and nano time costs 30% on windows. On linux, the cost is higher due to differences in how it's implemented. On Sun, Apr 3, 2011 at 5:28 AM, sebb <[email protected]> wrote: > On 3 April 2011 08:32, Ben Cuthbert <[email protected]> wrote: >> I see the nanotime. But the time in the sampler results is reported in ms. >> So when you have you data >> it just says 0. I would like it to go one further and report a low level. > > Sorry, that's not possible currently. > > Changing the elapsed time to nanoSecs would break compatibility, and > such a level of precision is illusory anyway for almost all of the > samplers. > > It might be possible to keep a separate field for nanoSeconds and > report that, but I'm not sure there's sufficient need to warrant the > change and additional data storage. > >> Regards >> On 2 Apr 2011, at 16:18, sebb wrote: >> >>> On 31 March 2011 19:41, Ben Cuthbert <[email protected]> wrote: >>>> All >>>> >>>> I have been looking over the code in the JUnitSampler code under the >>>> jmeter source. >>>> I would like to make a change to use nanoTime() rather than milliseconds. >>> >>> Why? >>> >>>> I can see in the AnnotatedTestCase there is an elapsed time. But I can't >>>> see how it >>>> is returned to a results table. Any ideas? >>> >>> The time in AnnotatedTestCase is only used for reporting timeout errors. >>> >>> The actual sample time is calculated using SampleResult.sampleStart() >>> and sampleEnd() which already use nanoTime(). >>> >>> --------------------------------------------------------------------- >>> To unsubscribe, e-mail: [email protected] >>> For additional commands, e-mail: [email protected] >>> >> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: [email protected] >> For additional commands, e-mail: [email protected] >> >> > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
