4 ms·
that would work on x86 but not on an arch that can re-order loads (e.g. arm)
by blibble 3y ago
that would work on x86 but not on an arch that can re-order loads (e.g. arm)
- insanitybit 3y agoI'm assuming it would be a fenced operation
- blibble 3y agobut then there's still an advantage to using the fenceless version :) (admittedly it's python so it's so slow it's probably not even measurable)
- insanitybit 3y agoYeah, for Python I feel like the difference between fenced vs unfenced doesn't matter. The primary cost is around your L3 cache getting slammed with contentious atomics but your L3 is already absolutely fucked if you're using Python.