Alexander Lakhin <exclusion@gmail.com> writes:
> 24.09.2025 13:45, Michael Banck wrote:
>> How much timer resolution do we require from the system? GNU Mach seems
>> to (at least try to) guarantee that the timer won't go backwards, but it
>> does not guarantee (currently) that two consecutive clock_gettime()
>> calls will return something different in all cases.
> Regarding the lowest timer resolution, as I mentioned at [3], 32k_counter
> gives only 0.030517 sec...
We are currently doing a short pg_test_timing run in every BF run,
but with only a cursory regex-based sanity check on the output.
Since it's a TAP test, we could easily report the full output in
the TAP log without causing problems. I was already thinking about
doing that, and if there's some question about the minimum expected
timer resolution then it's really silly to not be capturing that
data.
I will go do that, and in a few day's time we should have enough
reports to see what we can realistically expect.
regards, tom lane
