You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As I understand it, the time measurements in hprof are reset at the start of each frame. print_timing() will print the timings of just the last frame.
I think it could be useful to measure timings for a longer duration (e.g. 1 second), optionally. For one thing, this would possibly smooth out random fluctuations in the timings.
Secondly, in my use case, I have some functions that are called periodically (e.g. starting a new game tick 20 times a second), but not necessarily every frame. So in the print_timing() output, I'll get either 0% or the percentage of time in the last frame for these periodical functions. Here, I think it would be useful to instead get the time taken in e.g. the last second.
The text was updated successfully, but these errors were encountered:
In my case, also, what I care about is total time spent under a node, over all frames, not just the last frame. In fact, I had implicitly assumed this is what the crate does and was surprised at all those "1"s in the call counts.
As I understand it, the time measurements in hprof are reset at the start of each frame.
print_timing()
will print the timings of just the last frame.I think it could be useful to measure timings for a longer duration (e.g. 1 second), optionally. For one thing, this would possibly smooth out random fluctuations in the timings.
Secondly, in my use case, I have some functions that are called periodically (e.g. starting a new game tick 20 times a second), but not necessarily every frame. So in the
print_timing()
output, I'll get either0%
or the percentage of time in the last frame for these periodical functions. Here, I think it would be useful to instead get the time taken in e.g. the last second.The text was updated successfully, but these errors were encountered: