fix: fix gpu resource spec in llm deployments (#1812)

examples/llm/deploy/agg.yaml

@@ -79,11 +79,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest

examples/llm/deploy/agg_router.yaml

@@ -104,11 +104,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest

examples/llm/deploy/disagg.yaml

@@ -79,11 +79,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest
@@ -106,11 +106,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest

examples/llm/deploy/disagg_router.yaml

@@ -104,11 +104,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest
@@ -131,11 +131,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest

examples/vllm_v0/deploy/agg.yaml

@@ -54,11 +54,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest

examples/vllm_v0/deploy/disagg.yaml

@@ -54,11 +54,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest
@@ -81,11 +81,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest

examples/vllm_v0/deploy/disagg_planner.yaml

@@ -55,11 +55,11 @@ spec:
         requests:
           cpu: "20"
           memory: "40Gi"
-          nvidia.com/gpu: "2"
+          gpu: "2"
         limits:
           cpu: "20"
           memory: "40Gi"
-          nvidia.com/gpu: "2"
+          gpu: "2"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest
@@ -83,11 +83,11 @@ spec:
         requests:
           cpu: "20"
           memory: "40Gi"
-          nvidia.com/gpu: "2"
+          gpu: "2"
         limits:
           cpu: "20"
           memory: "40Gi"
-          nvidia.com/gpu: "2"
+          gpu: "2"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest

examples/vllm_v1/deploy/agg.yaml

@@ -79,11 +79,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest

examples/vllm_v1/deploy/disagg.yaml

@@ -79,11 +79,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest
@@ -106,11 +106,11 @@ spec:
         requests:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
         limits:
           cpu: "10"
           memory: "20Gi"
-          nvidia.com/gpu: "1"
+          gpu: "1"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest

examples/vllm_v1/deploy/disagg_planner.yaml

@@ -81,11 +81,11 @@ spec:
         requests:
           cpu: "20"
           memory: "40Gi"
-          nvidia.com/gpu: "2"
+          gpu: "2"
         limits:
           cpu: "20"
           memory: "40Gi"
-          nvidia.com/gpu: "2"
+          gpu: "2"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest
@@ -109,11 +109,11 @@ spec:
         requests:
           cpu: "20"
           memory: "40Gi"
-          nvidia.com/gpu: "2"
+          gpu: "2"
         limits:
           cpu: "20"
           memory: "40Gi"
-          nvidia.com/gpu: "2"
+          gpu: "2"
       extraPodSpec:
         mainContainer:
           image: nvcr.io/nvidia/ai-dynamo/vllm-runtime:latest