Releases: NVIDIA/gpu-operator
Releases · NVIDIA/gpu-operator
GPU Operator 1.10.1 Release
GPU Operator 1.10.0 Release
GPU Operator 1.9.1 Release
GPU Operator 1.9.0 Release
GPU Operator 1.9.0-beta Release
New Features
- Support for NVIDIA Data Center GPU Driver version
470.82.01. - MIG Manager support with Preinstalled GPU Drivers on the node.
- Entitlement free driver builds using DriverToolkit images on RedHat OpenShift 4.9.
- Support for enabling GPUDirect RDMA with Preinstalled MOFED drivers on the node using
driver.rdma.useHostMofedoption. - Upgrade support of GPU Operator and Operand Pods using Operator Lifecycle Manager(OLM) on RedHat OpenShift.
Improvements
- Automatic detection of default runtime used in the cluster and deprecate
operator.defaultRuntimeparameter. - Installation of GPU Operator and its Operands into a single user specified Namespace.
- Automatic detection and unload of
nouveaudriver prior to installing NVIDIA GPU Driver on each node. - Option to mount a ConfigMap of self-signed certificates into driver container using
driver.certConfig.nameparameter. This can be
used for SSL connections to Proxy server or private package repositories.
Fixed issues
- DCGM Exporter in
CrashLoopBackOffas it cannot connect to DCGM port on the same node.
Known Limitations
- Upgrades from
v1.8.xtov1.9.xare not supported due to change in default Namespacegpu-operator-resourcesused earlier.
Starting fromv1.9.xuser can specify custom Namespace for both GPU Operator and Operand Pods to be installed. User would need
to un-install old version prior to installingv1.9.xversion of GPU Operator. - Collection of GPU metrics in MIG mode is not supported with
470+drivers.