EventList

class tensorplay.profiler.EventList(raw=(), base_ns=0, gpu_activities=None, mem_events=None, samples=None, with_stack=False)[source]

Raw profiler events with analysis and export helpers.

copy()

Return a shallow copy of the list.

count(value, /)

Return number of occurrences of value.

device_kernels(sort_by=None)[source]

What the device ran, by the name the device knows it by.

Separate from the per-operation view because a kernel is not an operation: work this project’s own code launches belongs to whatever operation happened to be in flight, and naming it is the only way to find out it ran at all.

property device_time_by_span

For each collected span, the device time of the work inside it.

Kept alongside the events because the two answer different questions: a span’s own event says how long the span took on the host, and this says how long the device was busy for it, which for a span that only encloses work is the whole of what was being asked.

index(value, start=0, stop=9223372036854775807, /)

Return first index of value.

Raises ValueError if the value is not present.

pop(index=-1, /)

Remove and return item at index (default last).

Raises IndexError if list is empty or index is out of range.

remove(value, /)

Remove first occurrence of value.

Raises ValueError if the value is not present.

reverse()

Reverse IN PLACE.

sort(*, key=None, reverse=False)

Sort the list in ascending order and return None.

The sort is in-place (i.e. the list itself is modified) and stable (i.e. the order of two equal elements is maintained).

If a key function is given, apply it once to each list item and sort them, ascending or descending, according to their function values.

The reverse flag can be set to sort in descending order.