feat: add custom gpu type to CRD (#4408)

deploy/cloud/operator/internal/dynamo/graph.go

@@ -34,7 +34,6 @@ import (
 	"k8s.io/utils/ptr"
 
 	grovev1alpha1 "github.com/NVIDIA/grove/operator/api/core/v1alpha1"
-	"github.com/ai-dynamo/dynamo/deploy/cloud/operator/api/dynamo/common"
 	"github.com/ai-dynamo/dynamo/deploy/cloud/operator/api/v1alpha1"
 	commonconsts "github.com/ai-dynamo/dynamo/deploy/cloud/operator/internal/consts"
 	"github.com/ai-dynamo/dynamo/deploy/cloud/operator/internal/controller_common"
@@ -63,15 +62,15 @@ type Autoscaling struct {
 }
 
 type Config struct {
-	Dynamo       *DynamoConfig        `yaml:"dynamo,omitempty"`
-	Resources    *Resources           `yaml:"resources,omitempty"`
-	Traffic      *Traffic             `yaml:"traffic,omitempty"`
-	Autoscaling  *Autoscaling         `yaml:"autoscaling,omitempty"`
-	HttpExposed  bool                 `yaml:"http_exposed,omitempty"`
-	ApiEndpoints []string             `yaml:"api_endpoints,omitempty"`
-	Workers      *int32               `yaml:"workers,omitempty"`
-	TotalGpus    *int32               `yaml:"total_gpus,omitempty"`
-	ExtraPodSpec *common.ExtraPodSpec `yaml:"extraPodSpec,omitempty"`
+	Dynamo       *DynamoConfig          `yaml:"dynamo,omitempty"`
+	Resources    *Resources             `yaml:"resources,omitempty"`
+	Traffic      *Traffic               `yaml:"traffic,omitempty"`
+	Autoscaling  *Autoscaling           `yaml:"autoscaling,omitempty"`
+	HttpExposed  bool                   `yaml:"http_exposed,omitempty"`
+	ApiEndpoints []string               `yaml:"api_endpoints,omitempty"`
+	Workers      *int32                 `yaml:"workers,omitempty"`
+	TotalGpus    *int32                 `yaml:"total_gpus,omitempty"`
+	ExtraPodSpec *v1alpha1.ExtraPodSpec `yaml:"extraPodSpec,omitempty"`
 }
 
 type ServiceConfig struct {
@@ -150,7 +149,7 @@ func GenerateDynamoComponentsDeployments(ctx context.Context, parentDynamoGraphD
 		if component.ComponentType == commonconsts.ComponentTypePlanner {
 			// ensure that the extraPodSpec is not nil
 			if deployment.Spec.ExtraPodSpec == nil {
-				deployment.Spec.ExtraPodSpec = &common.ExtraPodSpec{}
+				deployment.Spec.ExtraPodSpec = &v1alpha1.ExtraPodSpec{}
 			}
 			// ensure that the embedded PodSpec struct is not nil
 			if deployment.Spec.ExtraPodSpec.PodSpec == nil {
@@ -231,10 +230,10 @@ func overrideWithDynDeploymentConfig(ctx context.Context, dynamoDeploymentCompon
 			dynamoDeploymentComponent.Spec.Replicas = componentDynConfig.ServiceArgs.Workers
 		}
 		if componentDynConfig.ServiceArgs != nil && componentDynConfig.ServiceArgs.Resources != nil {
-			requests := &common.ResourceItem{}
-			limits := &common.ResourceItem{}
+			requests := &v1alpha1.ResourceItem{}
+			limits := &v1alpha1.ResourceItem{}
 			if dynamoDeploymentComponent.Spec.Resources == nil {
-				dynamoDeploymentComponent.Spec.Resources = &common.Resources{
+				dynamoDeploymentComponent.Spec.Resources = &v1alpha1.Resources{
 					Requests: requests,
 					Limits:   limits,
 				}

deploy/helm/chart/templates/deployment.yaml

@@ -66,29 +66,47 @@ spec:
         - -m
         - dynamo.frontend
         {{- end }}
-        {{ if $serviceSpec.resources }}
+        {{- if $serviceSpec.resources }}
+        {{- $hasResources := false }}
+        {{- if or $serviceSpec.resources.requests $serviceSpec.resources.limits }}
+        {{- $hasResources = true }}
+        {{- end }}
+        {{- if $hasResources }}
         resources:
+          {{- if $serviceSpec.resources.requests }}
+          {{- $requestsGpuResourceName := "nvidia.com/gpu" }}
+          {{- if $serviceSpec.resources.requests.gpuType }}
+          {{- $requestsGpuResourceName = $serviceSpec.resources.requests.gpuType }}
+          {{- end }}
           requests:
-            {{ if $serviceSpec.resources.cpu }}
-            cpu: "{{ $serviceSpec.resources.cpu }}"
-            {{ end }}
-            {{ if $serviceSpec.resources.memory }}
-            memory: "{{ $serviceSpec.resources.memory }}"
-            {{ end }}
-            {{ if $serviceSpec.resources.gpu }}
-            nvidia.com/gpu: "{{ $serviceSpec.resources.gpu }}"
-            {{ end }}
+            {{- if $serviceSpec.resources.requests.cpu }}
+            cpu: "{{ $serviceSpec.resources.requests.cpu }}"
+            {{- end }}
+            {{- if $serviceSpec.resources.requests.memory }}
+            memory: "{{ $serviceSpec.resources.requests.memory }}"
+            {{- end }}
+            {{- if $serviceSpec.resources.requests.gpu }}
+            {{ $requestsGpuResourceName }}: "{{ $serviceSpec.resources.requests.gpu }}"
+            {{- end }}
+          {{- end }}
+          {{- if $serviceSpec.resources.limits }}
+          {{- $limitsGpuResourceName := "nvidia.com/gpu" }}
+          {{- if $serviceSpec.resources.limits.gpuType }}
+          {{- $limitsGpuResourceName = $serviceSpec.resources.limits.gpuType }}
+          {{- end }}
           limits:
-            {{ if $serviceSpec.resources.cpu }}
-            cpu: "{{ $serviceSpec.resources.cpu }}"
-            {{ end }}
-            {{ if $serviceSpec.resources.memory }}
-            memory: "{{ $serviceSpec.resources.memory }}"
-            {{ end }}
-            {{ if $serviceSpec.resources.gpu }}
-            nvidia.com/gpu: "{{ $serviceSpec.resources.gpu }}"
-            {{ end }}
-        {{ end }}
+            {{- if $serviceSpec.resources.limits.cpu }}
+            cpu: "{{ $serviceSpec.resources.limits.cpu }}"
+            {{- end }}
+            {{- if $serviceSpec.resources.limits.memory }}
+            memory: "{{ $serviceSpec.resources.limits.memory }}"
+            {{- end }}
+            {{- if $serviceSpec.resources.limits.gpu }}
+            {{ $limitsGpuResourceName }}: "{{ $serviceSpec.resources.limits.gpu }}"
+            {{- end }}
+          {{- end }}
+        {{- end }}
+        {{- end }}
         {{- if $serviceSpec.envFromSecret }}
         envFrom:
         - secretRef:

deploy/helm/chart/templates/grove-podgangset.yaml

@@ -39,27 +39,45 @@ spec:
           - name: main
             image: {{ $serviceSpec.extraPodSpec.mainContainer.image }}
             {{- if $serviceSpec.resources }}
+            {{- $hasResources := false }}
+            {{- if or $serviceSpec.resources.requests $serviceSpec.resources.limits }}
+            {{- $hasResources = true }}
+            {{- end }}
+            {{- if $hasResources }}
             resources:
+              {{- if $serviceSpec.resources.requests }}
+              {{- $requestsGpuResourceName := "nvidia.com/gpu" }}
+              {{- if $serviceSpec.resources.requests.gpuType }}
+              {{- $requestsGpuResourceName = $serviceSpec.resources.requests.gpuType }}
+              {{- end }}
               requests:
-                {{- if $serviceSpec.resources.cpu }}
-                cpu: "{{ $serviceSpec.resources.cpu }}"
+                {{- if $serviceSpec.resources.requests.cpu }}
+                cpu: "{{ $serviceSpec.resources.requests.cpu }}"
                 {{- end }}
-                {{- if $serviceSpec.resources.memory }}
-                memory: "{{ $serviceSpec.resources.memory }}"
+                {{- if $serviceSpec.resources.requests.memory }}
+                memory: "{{ $serviceSpec.resources.requests.memory }}"
                 {{- end }}
-                {{- if $serviceSpec.resources.gpu }}
-                nvidia.com/gpu: "{{ $serviceSpec.resources.gpu }}"
+                {{- if $serviceSpec.resources.requests.gpu }}
+                {{ $requestsGpuResourceName }}: "{{ $serviceSpec.resources.requests.gpu }}"
                 {{- end }}
+              {{- end }}
+              {{- if $serviceSpec.resources.limits }}
+              {{- $limitsGpuResourceName := "nvidia.com/gpu" }}
+              {{- if $serviceSpec.resources.limits.gpuType }}
+              {{- $limitsGpuResourceName = $serviceSpec.resources.limits.gpuType }}
+              {{- end }}
               limits:
-                {{- if $serviceSpec.resources.cpu }}
-                cpu: "{{ $serviceSpec.resources.cpu }}"
+                {{- if $serviceSpec.resources.limits.cpu }}
+                cpu: "{{ $serviceSpec.resources.limits.cpu }}"
                 {{- end }}
-                {{- if $serviceSpec.resources.memory }}
-                memory: "{{ $serviceSpec.resources.memory }}"
+                {{- if $serviceSpec.resources.limits.memory }}
+                memory: "{{ $serviceSpec.resources.limits.memory }}"
                 {{- end }}
-                {{- if $serviceSpec.resources.gpu }}
-                nvidia.com/gpu: "{{ $serviceSpec.resources.gpu }}"
+                {{- if $serviceSpec.resources.limits.gpu }}
+                {{ $limitsGpuResourceName }}: "{{ $serviceSpec.resources.limits.gpu }}"
                 {{- end }}
+              {{- end }}
+            {{- end }}
             {{- end }}
             workingDir: {{ $serviceSpec.extraPodSpec.mainContainer.workingDir }}
             {{- if $serviceSpec.extraPodSpec.mainContainer.command }}

docs/kubernetes/api_reference.md

@@ -123,6 +123,8 @@ _Appears in:_
 | `created` _boolean_ | Created indicates whether the DGD has been successfully created.<br />Used to prevent recreation if the DGD is manually deleted by users. |  |  |
 
 
+
+
 #### DynamoComponentDeployment
 
 
@@ -423,6 +425,41 @@ _Appears in:_
 | `ready` _boolean_ | Ready indicates whether the endpoint is ready to serve traffic<br />For LoRA models: true if the POST /loras request succeeded with a 2xx status code<br />For base models: always false (no probing performed) |  |  |
 
 
+#### ExtraPodMetadata
+
+
+
+
+
+
+
+_Appears in:_
+- [DynamoComponentDeploymentSharedSpec](#dynamocomponentdeploymentsharedspec)
+- [DynamoComponentDeploymentSpec](#dynamocomponentdeploymentspec)
+
+| Field | Description | Default | Validation |
+| --- | --- | --- | --- |
+| `annotations` _object (keys:string, values:string)_ |  |  |  |
+| `labels` _object (keys:string, values:string)_ |  |  |  |
+
+
+#### ExtraPodSpec
+
+
+
+
+
+
+
+_Appears in:_
+- [DynamoComponentDeploymentSharedSpec](#dynamocomponentdeploymentsharedspec)
+- [DynamoComponentDeploymentSpec](#dynamocomponentdeploymentspec)
+
+| Field | Description | Default | Validation |
+| --- | --- | --- | --- |
+| `mainContainer` _[Container](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.28/#container-v1-core)_ |  |  |  |
+
+
 #### IngressSpec
 
 
@@ -465,6 +502,8 @@ _Appears in:_
 | `secretName` _string_ | SecretName is the name of a Kubernetes Secret containing the TLS certificate and key. |  |  |
 
 
+
+
 #### ModelReference
 
 
@@ -556,6 +595,46 @@ _Appears in:_
 | `profilerImage` _string_ | ProfilerImage specifies the container image to use for profiling jobs.<br />This image contains the profiler code and dependencies needed for SLA-based profiling.<br />Example: "nvcr.io/nvidia/ai-dynamo/vllm-runtime:0.6.1" |  | Required: \{\} <br /> |
 
 
+#### ResourceItem
+
+
+
+
+
+
+
+_Appears in:_
+- [Resources](#resources)
+
+| Field | Description | Default | Validation |
+| --- | --- | --- | --- |
+| `cpu` _string_ | CPU specifies the CPU resource request/limit (e.g., "1000m", "2") |  |  |
+| `memory` _string_ | Memory specifies the memory resource request/limit (e.g., "4Gi", "8Gi") |  |  |
+| `gpu` _string_ | GPU indicates the number of GPUs to request.<br />Total number of GPUs is NumberOfNodes * GPU in case of multinode deployment. |  |  |
+| `gpuType` _string_ | GPUType can specify a custom GPU type, e.g. "gpu.intel.com/xe"<br />By default if not specified, the GPU type is "nvidia.com/gpu" |  |  |
+| `custom` _object (keys:string, values:string)_ | Custom specifies additional custom resource requests/limits |  |  |
+
+
+#### Resources
+
+
+
+Resources defines requested and limits for a component, including CPU, memory,
+GPUs/devices, and any runtime-specific resources.
+
+
+
+_Appears in:_
+- [DynamoComponentDeploymentSharedSpec](#dynamocomponentdeploymentsharedspec)
+- [DynamoComponentDeploymentSpec](#dynamocomponentdeploymentspec)
+
+| Field | Description | Default | Validation |
+| --- | --- | --- | --- |
+| `requests` _[ResourceItem](#resourceitem)_ | Requests specifies the minimum resources required by the component |  |  |
+| `limits` _[ResourceItem](#resourceitem)_ | Limits specifies the maximum resources allowed for the component |  |  |
+| `claims` _[ResourceClaim](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.28/#resourceclaim-v1-core) array_ | Claims specifies resource claims for dynamic resource allocation |  |  |
+
+
 #### SharedMemorySpec