3 ms·
I am not sure they are weighting wakeups heavily. "but with a “tax” of 500 microseconds for each wakeup" could be a reasonable estimate of the time the OS needs
by Someone 11y ago
I am not sure they are weighting wakeups heavily. "but with a “tax” of 500 microseconds for each wakeup" could be a reasonable estimate of the time the OS needs after wakeup before it can hand over the CPU to the process that did the wake up. That time isn't spend in the process that wakes up the system, but can be attributed to it.
- nnethercote 11y agoThe article discusses this. Look again at the first table. Compare the power values (measured in Watts) which climb slowly, vs. the "Energy Impact" values which climb incredibly quickly. If "Energy Impact" was a good measure it would correlate closely with real power consumption. It doesn't.