On Thursday, 22 May 2025 22:58:37 Central Daylight Time Igor Mironchik wrote: > PASS : Test::initTestCase() > PASS : Test::test_func() > RESULT : Test::test_func(): > 0.000035 msecs per iteration (total: 75, iterations: 2097152) > PASS : Test::test_template() > RESULT : Test::test_template(): > 0.0000045 msecs per iteration (total: 76, iterations: 16777216) > PASS : Test::cleanupTestCase() > > This is a whole order of magnitude.
Use -perf to benchmark tiny optimisations like that. Wall-clock time is not a good measure when the CPU frequency can change. The -iterations counter helps if the result is too small. -- Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org Principal Engineer - Intel Platform & System Engineering