undefined | Better HN

0 pointsvardump1y ago0 comments

I think it’s mostly due to Tracy’s poor timing calibration code. TSC is good accuracy and latency wise.

0 comments

The problem is not the timestamping, it's the queue used to push profiling events which is not fast enough

That queue is about as fast as it gets, <10 ns. Timestamp is taken before queueing.

Again, due to bad calibration code the measured timestamps have quite a bit jitter.

Edit: TSC might not be synchronized in multi-socket systems. (Multiple physical CPU sockets). That can generate a large error.

j / k navigate · click thread line to collapse

The problem is not the timestamping, it's the queue used to push profiling events which is not fast enough

vardumpOP1y ago

That queue is about as fast as it gets, <10 ns. Timestamp is taken before queueing.

Again, due to bad calibration code the measured timestamps have quite a bit jitter.

Edit: TSC might not be synchronized in multi-socket systems. (Multiple physical CPU sockets). That can generate a large error.

j / k navigate · click thread line to collapse