Interpreting average value in throughput result

[Aleksey Shipilev]

That is the measurement error. The "?" sign there is actually "±".

So the pattern is $score ± $score_error $unit [$aggregation_type].

-Aleksey.

On 11/05/2014 07:20 PM, Vladimir Dzhuvinov wrote:
> Thanks for the quick reply.
> 
>> Result: 327.893 ?(99.9%) 84.101 ops/s [Average]
> 
> If 327.893 is the average, then what is the second value (84.101)?
> 
> Pardon me if I have missed something obvious.
> 
> Vladimir
> 
> 
> On 5.11.2014 18:14, Aleksey Shipilev wrote:
>> Hi Vladimir,
>>
>> That is an average across multiple iterations. In your case, the average
>> across 5 values we got from "Iteration \d+". There are cases where we
>> compute the Max, Min or something else, notably in profiler metrics.
>>
>> -Aleksey.
>>
>> On 11/05/2014 07:11 PM, Vladimir Dzhuvinov wrote:
>>> Hello guys,
>>>
>>> How should I interpret the "84.101 ops/s [Average]" value from the
>>> output below? How does it relate to the first result value (327.893)?
>>>
>>> Thanks,
>>>
>>> Vladimir
>>>
>>>
>>> ***
>>> java -jar benchmarks.jar -wi 2 -i 5 -f 0 -t 25
>>>
>>> # Run progress: 0.00% complete, ETA 01:30:36
>>> # Warmup: 2 iterations, 1 s each
>>> # Measurement: 5 iterations, 60 s each
>>> # Timeout: 10 min per iteration
>>> # Threads: 25 threads, will synchronize iterations
>>> # Benchmark mode: Throughput, ops/time
>>> # Benchmark:
>>> com.nimbusds.openid.connect.provider.benchmarks.AuthorizationCodeGrantBenchmark.authorize
>>> # Fork: N/A, test runs in the existing VM
>>> # Warmup Iteration   1: 143.927 ops/s
>>> # Warmup Iteration   2: 157.111 ops/s
>>> Iteration   1: 296.617 ops/s
>>> Iteration   2: 339.026 ops/s
>>> Iteration   3: 346.989 ops/s
>>> Iteration   4: 313.571 ops/s
>>> Iteration   5: 343.262 ops/s
>>>
>>> Result: 327.893 ?(99.9%) 84.101 ops/s [Average]
>>>   Statistics: (min, avg, max) = (296.617, 327.893, 346.989), stdev = 21.841
>>>   Confidence interval (99.9%): [243.792, 411.994]
>>>
>>
>