odd timing behaviour on panda
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Linaro Ubuntu |
Fix Released
|
Undecided
|
Unassigned | ||
linaro-landing-team-ti |
Fix Released
|
Low
|
David Long |
Bug Description
I've got a set of benchmarks that use clock_gettime like:
clock_
for(
dostuff
}
clock_
nsdiff=
nsdiff+
and I've just reinstalled our local panda to using Linaro 11.09 (kernel 3.0.0-1404-
and it's starting to get weird timing artifacts.
For example:
smarter_
smarter_
smarter_
Now there is no way that loop took 0.000000 ns! Running the same binary on an older installation runs fine.
Dave
affects: | linux-linaro-omap (Ubuntu) → ubuntu-leb |
affects: | ubuntu-leb → linaro-ubuntu |
tags: | added: panda |
Changed in linaro-landing-team-ti: | |
importance: | Undecided → Medium |
assignee: | nobody → warmcat (andy-warmcat) |
milestone: | none → 2011.11 |
Changed in linaro-landing-team-ti: | |
assignee: | warmcat (andy-warmcat) → David Long (dave-long) |
Changed in linaro-ubuntu: | |
assignee: | nobody → Ricardo Salveti (rsalveti) |
Changed in linaro-landing-team-ti: | |
importance: | Undecided → Low |
Changed in linaro-landing-team-ti: | |
status: | Confirmed → Fix Released |
Changed in linaro-ubuntu: | |
assignee: | Ricardo Salveti (rsalveti) → nobody |
status: | Confirmed → Fix Released |
If there is some granularity to the timing it doesn't sound crazy you read the same time twice sometimes. There's a 32kHz clock I know is used for some things, that would have high granularity.
Is this gettimeofday() doing this too?