[Question] Measuring page-faults
Is there any example of how to measure page-faults associated with a Level-Zero buffer with Level Zero?
I guess I would need this data structure: zet_debug_event_info_page_fault_t but I can't find how to query this.
https://spec.oneapi.io/level-zero/latest/tools/api.html#zet-debug-event-info-page-fault-t
I am looking for something similar to NVIDIA nvprof and "GPU page fault groups" but for Intel GPUs with Level Zero.
@jjfumero are you asking about measuring the overhead of a pagefault, to understand the performance?
I would like to get the number of pagefaults for a given execution. Also, the overhead of page fault in Level Zero would be useful. is it possible to get this with the current API?
thanks for the clarification @jjfumero . At this moment we dont have that interface because we dont know of support in L0 drivers to be able to implement such feature. Of course, that could change in the future.