You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I made a small test to understand the NBench infrastructure, but the results are quite surprising.
The SkipWarmups is set to true, the NumberOfIterations is set to 2, but the SetUp is called three times.
After the first SetUp call, the SleepTroughput method is called 80 times - I assume these calls are the estimation calls.
The second and third SetUp are the actual benchmark calls, but there are only 10 SleepTroughput calls after them. And the counter's actual value is 9.59 operations
In my opinion the expected outcome should be around 20 operations (sleep for 100 ms in 2000 ms time frame).
If RunTimeMilliseconds is 1000, then the only change is that the first SetUp call is continued by 40 SleepTroughput calls instead of the previous 80, but the number of operations during the benchmark test is the same (~10).
I made a small test to understand the NBench infrastructure, but the results are quite surprising.
The SkipWarmups is set to true, the NumberOfIterations is set to 2, but the SetUp is called three times.
After the first SetUp call, the SleepTroughput method is called 80 times - I assume these calls are the estimation calls.
The second and third SetUp are the actual benchmark calls, but there are only 10 SleepTroughput calls after them. And the counter's actual value is 9.59 operations
In my opinion the expected outcome should be around 20 operations (sleep for 100 ms in 2000 ms time frame).
If RunTimeMilliseconds is 1000, then the only change is that the first SetUp call is continued by 40 SleepTroughput calls instead of the previous 80, but the number of operations during the benchmark test is the same (~10).
The text was updated successfully, but these errors were encountered: