cuda.bindings.nvml.ObservedMetrics#

class cuda.bindings.nvml.ObservedMetrics(size=1)#

Empty-initialize an array of nvmlObservedMetrics_t. The resulting object is of length size and of dtype observed_metrics_dtype. If default-constructed, the instance represents a single struct.

Parameters:

size (int) – number of structs, default=1.

See also

nvmlObservedMetrics_t

__init__(*args, **kwargs)#

Methods

__init__(*args, **kwargs)

from_buffer(buffer)

Create an ObservedMetrics instance with the memory from the given buffer.

from_data(data)

Create an ObservedMetrics instance wrapping the given NumPy array.

from_ptr(intptr_t ptr, size_t size=1, ...[, ...])

Create an ObservedMetrics instance wrapping the given pointer.

Attributes

b_valid

non-zero if observed metrics are valid.

core_rail

Observed core rail metrics.

fb_rail

Observed fb rail metrics.

initial_dramclk_est

(array of length 3).Initial DRAMCLK estimates for different scenarios.

perf_metrics

Observed performance metrics.

ptr

Get the pointer address to the data as Python int.

tgp_pwr_tuple

Observed Total Graphics Power (TGP) in milliwatts.