Closed sloorush closed 8 months ago
False alarm: The counter does not report incorrect time but it reports the CPU time instead of the elapsed time. I am not sure why CPU time and elapsed time are so different. I think running the tests multiple times causes too much time spent on IO things?
Log:
]performance.runtime 'unittest.RunTests tests.unique 0 0 5'
* Benchmarking "unittest.RunTests tests.unique 0 0 5"
Options:
Version information:
Dyalog 18.2.45405 64-bit Unicode, BuildID 8ebc0eba
OS Windows 10 or Windows Server 2016 64-bit
Link 3.0.19
SALT 2.9
UCMD 2.51
.NET
WS 18.2
Repetition: 1 with Option:
⎕RL is set to: 1956353843
#.tests.unique.test_unique
Repetition 1 tests completed: 9488 ran, 0 failed test, 9488 successes
Repetition: 2 with Option:
⎕RL is set to: 1488808292
#.tests.unique.test_unique
Repetition 2 tests completed: 9488 ran, 0 failed test, 9488 successes
Repetition: 3 with Option:
⎕RL is set to: 406691272
#.tests.unique.test_unique
Repetition 3 tests completed: 9488 ran, 0 failed test, 9488 successes
Repetition: 4 with Option:
⎕RL is set to: 834366396
#.tests.unique.test_unique
Repetition 4 tests completed: 9488 ran, 0 failed test, 9488 successes
Repetition: 5 with Option:
⎕RL is set to: 1955270430
#.tests.unique.test_unique
Repetition 5 tests completed: 9488 ran, 0 failed test, 9488 successes
All tests completed: 5 repetitions, total tests: 47440 ran, 0 failed test, 47440 successes in 7.031 s
(ms)
CPU (avg): 7031
Elapsed: 12198
Originally posted by @sloorush in https://github.com/Dyalog/ullu/issues/45#issuecomment-1865916900