Skip to content

Commit

Permalink
intel-gpu-resource-driver: update to v0.5.1 version (#58)
Browse files Browse the repository at this point in the history 
* intel-gpu-resource-driver: update to v0.5.1 version
  • Loading branch information
@oxxenix
oxxenix authored Aug 8, 2024
1 parent 55ab2ee commit 41c316c
Show file tree
Hide file tree
Showing 11 changed files with 388 additions and 404 deletions.
4 changes: 2 additions & 2 deletions charts/intel-gpu-resource-driver/Chart.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,5 +3,5 @@ name: intel-gpu-resource-driver
description: A Helm chart for a Dynamic Resource Allocation (DRA) Intel GPU Resource Driver

type: application
version: 0.4.0
appVersion: "v0.4.0"
version: 0.5.1
appVersion: "v0.5.1"
12 changes: 9 additions & 3 deletions charts/intel-gpu-resource-driver/README.md
View file
Original file line number Diff line number Diff line change
@@ -1,5 +1,12 @@
# Dynamic Resource Allocation (DRA) Intel GPU Driver Helm Chart

## The chart installs GPU resource driver:

- [GPU](https://github.com/intel-innersource/containers.orchestrators.kubernetes.intel-resource-drivers-for-kubernetes/doc/gpu/README.md)

More info: [Intel Resource Drivers for Kubernetes](https://github.com/intel-innersource/containers.orchestrators.kubernetes.intel-resource-drivers-for-kubernetes)


## Get Helm Repository Info
```
helm repo add intel https://intel.github.io/helm-charts/
Expand All @@ -12,8 +19,7 @@ You can execute `helm search repo intel` command to see pulled charts [optional]
CRDs of the GPU driver are installed as part of the chart first.

```
helm install intel-gpu-resource-driver intel/intel-gpu-resource-driver \
--create-namespace --namespace intel-gpu-resource-driver
helm install intel-gpu-resource-driver intel/intel-gpu-resource-driver
```
## Upgrade Chart
```
Expand All @@ -39,7 +45,7 @@ You may also run `helm show values` on this chart's dependencies for additional
| image.repository | string | `intel` |
| image.name | string | `"intel-gpu-resource-driver"` |
| image.pullPolicy | string | `"IfNotPresent"` |
| image.tag | string | `"v0.4.0"` |
| image.tag | string | `"v0.5.1"` |

If you change the image tag to be used in Helm chart deployment, ensure that the version of the container image is consistent with CRDs and deployment YAMLs - they might change between releases.

Expand Down
330 changes: 173 additions & 157 deletions charts/intel-gpu-resource-driver/crds/gpu.resource.intel.com_gpuallocationstates.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@ apiVersion: apiextensions.k8s.io/v1
kind: CustomResourceDefinition
metadata:
annotations:
controller-gen.kubebuilder.io/version: v0.13.0
controller-gen.kubebuilder.io/version: v0.15.0
name: gpuallocationstates.gpu.resource.intel.com
spec:
group: gpu.resource.intel.com
Expand All @@ -14,162 +14,178 @@ spec:
singular: gas
scope: Namespaced
versions:
- name: v1alpha2
schema:
openAPIV3Schema:
description: GpuAllocationState holds the state required for allocation on
a node.
properties:
apiVersion:
description: 'APIVersion defines the versioned schema of this representation
of an object. Servers should convert recognized schemas to the latest
internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources'
type: string
kind:
description: 'Kind is a string value representing the REST resource this
object represents. Servers may infer this from the endpoint the client
submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds'
type: string
metadata:
type: object
spec:
description: GpuAllocationStateSpec is the spec for the GpuAllocationState
CRD.
properties:
allocatableDevices:
additionalProperties:
description: AllocatableGpu represents an allocatable Gpu on a node.
properties:
ecc:
description: True if ECC is enabled, might impact memory amount
and VF profiles.
type: boolean
maxvfs:
description: Greater than 0 if SR-IOV is supported / enabled.
format: int64
type: integer
memory:
description: Amount of local memory in MiB.
format: int64
maximum: 1048576
minimum: 0
type: integer
millicores:
description: Amount of GPU millicores.
format: int64
maximum: 1000
minimum: 0
type: integer
model:
description: pci-id of the Gpu device.
type: string
parentuid:
description: Device where VF should be / is provisioned.
type: string
type:
description: 'Type of the device: bare-metal Gpu or SR-IOV Virtual
Function (VF).'
enum:
- gpu
- vf
- any
type: string
uid:
description: 'Unique identifier of device: PCI address and PCI
Device ID.'
type: string
vfindex:
description: Index of SR-IOV Virtual Function
format: int64
type: integer
required:
- ecc
- maxvfs
- memory
- millicores
- model
- parentuid
- type
- uid
- vfindex
- name: v1alpha2
schema:
openAPIV3Schema:
description:
GpuAllocationState holds the state required for allocation on
a node.
properties:
apiVersion:
description: |-
APIVersion defines the versioned schema of this representation of an object.
Servers should convert recognized schemas to the latest internal value, and
may reject unrecognized values.
More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
type: string
kind:
description: |-
Kind is a string value representing the REST resource this object represents.
Servers may infer this from the endpoint the client submits requests to.
Cannot be updated.
In CamelCase.
More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
type: string
metadata:
type: object
spec:
description:
GpuAllocationStateSpec is the spec for the GpuAllocationState
CRD.
properties:
allocatableDevices:
additionalProperties:
description: AllocatableGpu represents an allocatable Gpu on a node.
properties:
ecc:
description:
True if ECC is enabled, might impact memory amount
and VF profiles.
type: boolean
maxvfs:
description: Greater than 0 if SR-IOV is supported / enabled.
format: int64
type: integer
memory:
description: Amount of local memory in MiB.
format: int64
maximum: 1048576
minimum: 0
type: integer
millicores:
description: Amount of GPU millicores.
format: int64
maximum: 1000
minimum: 0
type: integer
model:
description: pci-id of the Gpu device.
type: string
parentuid:
description: Device where VF should be / is provisioned.
type: string
type:
description:
"Type of the device: bare-metal Gpu or SR-IOV Virtual
Function (VF)."
enum:
- gpu
- vf
- any
type: string
uid:
description:
"Unique identifier of device: PCI address and PCI
Device ID."
type: string
vfindex:
description: Index of SR-IOV Virtual Function
format: int64
type: integer
required:
- ecc
- maxvfs
- memory
- millicores
- model
- parentuid
- type
- uid
- vfindex
type: object
type: object
type: object
allocatedClaims:
additionalProperties:
description: Resources that were allocated for the claim by controller.
properties:
gpus:
description: AllocatedGpus represents a list of allocated devices
on a node.
items:
description: AllocatedGpu represents an allocated Gpu on a
node.
properties:
memory:
description: Amount of local memory in MiB.
format: int64
maximum: 1048576
minimum: 0
type: integer
millicores:
description: Amount of GPU millicores.
format: int64
maximum: 1000
minimum: 0
type: integer
parentuid:
description: Device where VF should be / is provisioned.
type: string
profile:
description: Virtual Function profile defines amount of
local memory and time slice VF gets.
type: string
type:
description: 'Type of the device: bare-metal Gpu or SR-IOV
Virtual Function (VF).'
enum:
- gpu
- vf
- any
type: string
uid:
description: 'Unique identifier of device: PCI address
and PCI Device ID.'
type: string
vfindex:
description: Index of SR-IOV Virtual Function
format: int64
type: integer
required:
- memory
- millicores
- parentuid
- profile
- type
- uid
- vfindex
type: object
maxItems: 640
type: array
required:
- gpus
allocatedClaims:
additionalProperties:
description: Resources that were allocated for the claim by controller.
properties:
gpus:
description:
AllocatedGpus represents a list of allocated devices
on a node.
items:
description:
AllocatedGpu represents an allocated Gpu on a
node.
properties:
memory:
description: Amount of local memory in MiB.
format: int64
maximum: 1048576
minimum: 0
type: integer
millicores:
description: Amount of GPU millicores.
format: int64
maximum: 1000
minimum: 0
type: integer
parentuid:
description: Device where VF should be / is provisioned.
type: string
profile:
description:
Virtual Function profile defines amount of
local memory and time slice VF gets.
type: string
type:
description:
"Type of the device: bare-metal Gpu or SR-IOV
Virtual Function (VF)."
enum:
- gpu
- vf
- any
type: string
uid:
description:
"Unique identifier of device: PCI address
and PCI Device ID."
type: string
vfindex:
description: Index of SR-IOV Virtual Function
format: int64
type: integer
required:
- memory
- millicores
- parentuid
- profile
- type
- uid
- vfindex
type: object
maxItems: 640
type: array
required:
- gpus
type: object
type: object
type: object
taintedDevices:
additionalProperties:
description: TaintedGpu represents a tainted Gpu on a node.
properties:
reasons:
additionalProperties:
type: boolean
description: Reasons why device is tainted, which _all_ need
to be resolved, before device can be dropped from taints map.
type: object
taintedDevices:
additionalProperties:
description: TaintedGpu represents a tainted Gpu on a node.
properties:
reasons:
additionalProperties:
type: boolean
description: |-
Reasons why device is tainted, which _all_ need to be
resolved, before device can be dropped from taints map.
type: object
type: object
type: object
type: object
type: object
status:
type: string
type: object
served: true
storage: true
type: object
status:
type: string
type: object
served: true
storage: true
Loading

0 comments on commit 41c316c

Please sign in to comment.