MxNet unable to detect CPU on AWS/EMR/g4 instance

[patrick-nicolas]

The application works just fine on CPU. I followed the recommendations from issue 732
Issue-732
Which step am I missing?

GPU configuration:
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 450.172.01 Driver Version: 450.172.01 CUDA Version: 11.0 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 Tesla T4 Off | 00000000:00:1E.0 Off | 0 |
| N/A 33C P0 14W / 70W | 0MiB / 15109MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |

MxNet binary libraries
/home/hadoop/.djl.ai/mxnet/1.9.0-mkl-linux-x86_64/
├── libgfortran.so.3
├── libgomp.so.1
├── libmxnet.so
├── libopenblas.so.0
└── libquadmath.so.0

Relevant environment Variables
PATH: /home/hadoop/spark-3.3.0-bin-hadoop3-scala2.13/bin/:/home/hadoop/jdk-11/bin:/usr/local/cuda/bin:/usr/local/cuda/lib64:usr/local/cuda/targets/x86_64-linux/lib:/usr/lib64/qt-3.3/bin:/usr/local/bin:/usr/bin:/usr/local/sbin:/usr/sbin:/opt/aws/puppet/bin/
LD_LIBRARY_PATH: /usr/local/cuda/lib64:/usr/local/cuda/targets/x86_64-linux/lib
JAVA_HOME: /home/hadoop/jdk-11
SPARK_ENV_LOADED: 1
SPARK_SCALA_VERSION: 2.13
SPARK_HOME: /home/hadoop/spark-3.3.0-bin-hadoop3-scala2.13
SHELL: /bin/bash
USER: hadoop
SPARK_CONF_DIR: /home/hadoop/spark-3.3.0-bin-hadoop3-scala2.13/conf
HOME: /home/hadoop

-------------- Directories --------------
temp directory: /tmp
DJL cache directory: /home/hadoop/.djl.ai
Engine cache directory: /home/hadoop/.djl.ai

------------------ CUDA -----------------
GPU Count: 1
CUDA: 110
ARCH: 75
GPU(0) memory used: 106954752 bytes

----------------- Engines ---------------
MXNet version: 1.9.0
Default Engine: MXNet
Default Device: cpu()
MXNet: #0

[frankfliu]

@patrick-nicolas
MXNet 1.9.1 only support CUDA 11.2/10.2, see: https://docs.djl.ai/engines/mxnet/mxnet-engine/index.html#linux-gpu