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
After developing metrics on overall GPU utilization on macOS, I encountered a limitation in accessing the list of processes using the GPU and the GPU utilization (percentage) per process. Indeed, this information is not publicly available through any API. The only way to obtain it on macOS (as a user) is through the Activity Monitor.
This issue prompted me to consider other metrics that could be implemented for macOS, utilizing reliable information and public APIs. Here are a few ideas that came to my mind:
Network interfaces and related IP addresses (via struct ifaddrs).
Page ins, faults, and number of wakeups (all via struct rusage_info_v2).
A metric that displays the GPU utilization splitter by its two main components: Renderer Utilization % and Tiler Utilization %. I suggest structuring this metric in two colors, similar to memory usage.
The text was updated successfully, but these errors were encountered:
After developing metrics on overall GPU utilization on macOS, I encountered a limitation in accessing the list of processes using the GPU and the GPU utilization (percentage) per process. Indeed, this information is not publicly available through any API. The only way to obtain it on macOS (as a user) is through the Activity Monitor.
This issue prompted me to consider other metrics that could be implemented for macOS, utilizing reliable information and public APIs. Here are a few ideas that came to my mind:
Network interfaces
and relatedIP
addresses (viastruct ifaddrs
).Page ins
,faults
, and number ofwakeups
(all viastruct rusage_info_v2
).Renderer Utilization %
andTiler Utilization %
. I suggest structuring this metric in two colors, similar to memory usage.The text was updated successfully, but these errors were encountered: