Merge branch 'generic_nvidia_smi_err' into 'master'

adding a script to catch a generic nvidia-smi error

See merge request !587

roles/nhc/files/scripts/check_gpu_generic_err.nhc

+#!/bin/bash
+function check_gpu_generic_err() {
+#echo ">>> Checking for generic GPU errors >>>>>>>>>>>>>>>>>>>>>>>>>"
+MASSIVEDOC="Runs nvidia-smi to see if the GPUs have any errors. Errors may indicate faulty hardware"
+
+/usr/bin/nvidia-smi | ( ! grep ERR )
+ret=$?
+if [[ $ret -ne 0 ]]
+then
+        die 1 " $FUNCNAME ERROR GPU errors"
+        return 1
+else
+        return 0
+fi
+}