"Schock, Jonathan" <jonathan.schock(a)tum.de> writes:
I am becoming more and more sure, that the problem is
based on wrong
The kernel seems to run in correct order with the profiler not being
able to show that, because
the runtimes are too short. If I increase the runtime of the kernel
(e.g. by doubling the amount
of data to crunch on), the anomaly seems to vanish. I only ran into the
again if the kernel runtime is below ~500 µs.
I'm nevertheless interested in the reason, this is happening.
Happy to hear this, and I, too, would be curious to know what's causing
this. That said, if the timing you showed were right, then there'd be
guaranteed to be some incorrect results.
Thanks for following up!