Use nvidia-ml-py3 instead of nvidia-ml-py

README.rst

@@ -64,7 +64,7 @@ Optional dependencies:
 - ``kafka-python`` (for the Kafka export module)
 - ``matplotlib`` (for graphical/chart support)
 - ``netifaces`` (for the IP plugin)
-- ``nvidia-ml-py`` (for the GPU plugin) [Python 2-only]
+- ``nvidia-ml-py3`` (for the GPU plugin)
 - ``pika`` (for the RabbitMQ/ActiveMQ export module)
 - ``potsdb`` (for the OpenTSDB export module)
 - ``prometheus_client`` (for the Prometheus export module)

docs/aoa/gpu.rst

@@ -3,10 +3,8 @@
 GPU
 ===
 
-The GPU plugin is **only** compatible with ``NVIDIA`` GPUs.
-
 .. note::
-    You need to install the `nvidia-ml-py`_ library on your system.
+    You need to install the `nvidia-ml-py3`_ library on your system.
 
 The GPU stats are shown as a percentage of value and for the configured
 refresh time. The total GPU usage is displayed on the first line, the
@@ -47,4 +45,4 @@ GPU (PROC/MEM) Status
 ``>90%``       ``CRITICAL``
 ============== ============
 
-.. _nvidia-ml-py: https://pypi.python.org/pypi/nvidia-ml-py
+.. _nvidia-ml-py3: https://pypi.python.org/pypi/nvidia-ml-py3

optional-requirements.txt

@@ -8,7 +8,7 @@ influxdb
 kafka-python
 matplotlib
 netifaces
-nvidia-ml-py; python_version == "2.7"
+nvidia-ml-py3
 pika
 potsdb
 prometheus_client