refactor(srt): merge RDMA NIC fabric fields into rdma-devices

nccl-ib-hca, mooncake-devices and ib-devices held the same RDMA HCA list on
every cluster that set them. One Fabric field, rdma-devices, now carries that
list for NCCL_IB_HCA, the Mooncake device_name, SGLang
disaggregation-ib-device and IBDEVICES, and each cluster lists its HCAs once.

Bound recipes are unchanged. Rows whose recipes reference the field get new
fingerprints, since references are hashed as written.

Author: adibarra

inferencex-e2e/benchmarks/multi_node/srt-slurm-recipes/dsr1/sglang/mi355x-fp4/8k1k/disagg-variants.yaml

@@ -31,7 +31,7 @@ base:
         HF_HUB_CACHE: /hf_hub_cache/hub
         HUGGINGFACE_HUB_CACHE: /hf_hub_cache/hub
         PYTHONUNBUFFERED: '1'
-        IBDEVICES: '@fabric.ib-devices'
+        IBDEVICES: '@fabric.rdma-devices'
         GLOO_SOCKET_IFNAME: '@fabric.socket-ifname'
         NCCL_SOCKET_IFNAME: '@fabric.socket-ifname'
         MORI_RDMA_TC: '@fabric.mori-rdma-tc'
@@ -89,7 +89,7 @@ base:
         HF_HUB_CACHE: /hf_hub_cache/hub
         HUGGINGFACE_HUB_CACHE: /hf_hub_cache/hub
         PYTHONUNBUFFERED: '1'
-        IBDEVICES: '@fabric.ib-devices'
+        IBDEVICES: '@fabric.rdma-devices'
         GLOO_SOCKET_IFNAME: '@fabric.socket-ifname'
         NCCL_SOCKET_IFNAME: '@fabric.socket-ifname'
         MORI_RDMA_TC: '@fabric.mori-rdma-tc'

inferencex-e2e/benchmarks/multi_node/srt-slurm-recipes/dsr1/sglang/mi355x-fp8/8k1k/disagg-variants.yaml

@@ -31,7 +31,7 @@ base:
         HF_HUB_CACHE: /hf_hub_cache/hub
         HUGGINGFACE_HUB_CACHE: /hf_hub_cache/hub
         PYTHONUNBUFFERED: '1'
-        IBDEVICES: '@fabric.ib-devices'
+        IBDEVICES: '@fabric.rdma-devices'
         GLOO_SOCKET_IFNAME: '@fabric.socket-ifname'
         NCCL_SOCKET_IFNAME: '@fabric.socket-ifname'
         MORI_RDMA_TC: '@fabric.mori-rdma-tc'
@@ -84,7 +84,7 @@ base:
         HF_HUB_CACHE: /hf_hub_cache/hub
         HUGGINGFACE_HUB_CACHE: /hf_hub_cache/hub
         PYTHONUNBUFFERED: '1'
-        IBDEVICES: '@fabric.ib-devices'
+        IBDEVICES: '@fabric.rdma-devices'
         GLOO_SOCKET_IFNAME: '@fabric.socket-ifname'
         NCCL_SOCKET_IFNAME: '@fabric.socket-ifname'
         MORI_RDMA_TC: '@fabric.mori-rdma-tc'

inferencex-e2e/benchmarks/multi_node/srt-slurm-recipes/dsv4/sglang/b200-fp4/agentx/disagg-variants.yaml

@@ -103,7 +103,7 @@ base:
         disaggregation-transfer-backend: mooncake
         # Use the eight external HCAs. mlx5_6..9 are on a separate low-LID
         # fabric, while mlx5_bond_0 caused cross-fabric QP RTR timeouts.
-        disaggregation-ib-device: '@fabric.ib-devices'
+        disaggregation-ib-device: '@fabric.rdma-devices'
         disaggregation-mode: prefill
         load-balance-method: total_tokens
         mem-fraction-static: 0.85
@@ -175,7 +175,7 @@ base:
         enable-deepseek-v4-fp4-indexer: true
         enable-w4a4-mxfp4-megamoe: true
         disaggregation-transfer-backend: mooncake
-        disaggregation-ib-device: '@fabric.ib-devices'
+        disaggregation-ib-device: '@fabric.rdma-devices'
         disaggregation-mode: decode
         load-balance-method: total_tokens
         mem-fraction-static: 0.9

inferencex-e2e/benchmarks/multi_node/srt-slurm-recipes/dsv4/vllm/gb200-fp4/agentx/agg-mtp-variants.yaml

@@ -65,7 +65,7 @@ base:
         UCX_MEMTYPE_CACHE: "n"
         UCX_NET_DEVICES: "@fabric.ucx-net-devices"
         UCX_TLS: "rc,cuda_copy"
-        NCCL_IB_HCA: "@fabric.nccl-ib-hca"
+        NCCL_IB_HCA: "@fabric.rdma-devices"
       args:
         kv-events-config: '{"publisher":"zmq","topic":"kv-events","endpoint":"tcp://*:20080","enable_kv_cache_events":true}'
         served-model-name: deepseek-ai/DeepSeek-V4-Pro

inferencex-e2e/benchmarks/multi_node/srt-slurm-recipes/dsv4/vllm/gb200-fp4/agentx/agg-mtp3-variants.yaml

@@ -45,7 +45,7 @@ base:
           global_segment_size: "150GB"
           local_buffer_size: "4GB"
           protocol: "rdma"
-          device_name: "@fabric.mooncake-devices"
+          device_name: "@fabric.rdma-devices"
           mode: "embedded"
           enable_offload: false
   frontend:
@@ -93,7 +93,7 @@ base:
         UCX_MEMTYPE_CACHE: "n"
         UCX_NET_DEVICES: "@fabric.ucx-net-devices"
         UCX_TLS: "rc,cuda_copy"
-        NCCL_IB_HCA: "@fabric.nccl-ib-hca"
+        NCCL_IB_HCA: "@fabric.rdma-devices"
         NCCL_P2P_LEVEL: "NVL"
         MC_ENABLE_DEST_DEVICE_AFFINITY: "1"
         MC_STORE_CLIENT_METRIC: "1"

inferencex-e2e/benchmarks/multi_node/srt-slurm-recipes/dsv4/vllm/gb200-fp4/agentx/disagg-mtp-variants.yaml

@@ -62,7 +62,7 @@ base:
         UCX_MEMTYPE_CACHE: "n"
         UCX_NET_DEVICES: "@fabric.ucx-net-devices"
         UCX_TLS: "rc,cuda_copy"
-        NCCL_IB_HCA: "@fabric.nccl-ib-hca"
+        NCCL_IB_HCA: "@fabric.rdma-devices"
       args:
         kv-transfer-config: '{"kv_connector":"NixlConnector","kv_role":"kv_both"}'
         kv-events-config: '{"publisher":"zmq","topic":"kv-events","endpoint":"tcp://*:20080","enable_kv_cache_events":true}'
@@ -106,7 +106,7 @@ base:
         UCX_MEMTYPE_CACHE: "n"
         UCX_NET_DEVICES: "@fabric.ucx-net-devices"
         UCX_TLS: "rc,cuda_copy"
-        NCCL_IB_HCA: "@fabric.nccl-ib-hca"
+        NCCL_IB_HCA: "@fabric.rdma-devices"
       args:
         attention-config: '{"backend":"FLASHINFER_MLA_SPARSE_DSV4","use_fp4_indexer_cache":true}'
         enforce-eager: false

inferencex-e2e/benchmarks/multi_node/srt-slurm-recipes/dsv4/vllm/gb200-fp4/agentx/disagg-mtp3-variants.yaml

@@ -43,7 +43,7 @@ base:
           global_segment_size: "140GB"
           local_buffer_size: "4GB"
           protocol: "rdma"
-          device_name: "@fabric.mooncake-devices"
+          device_name: "@fabric.rdma-devices"
           mode: "embedded"
           enable_offload: false
   environment:
@@ -84,7 +84,7 @@ base:
         NCCL_CUMEM_ENABLE: "1"
         NCCL_MNNVL_ENABLE: "1"
         NCCL_NVLS_ENABLE: "1"
-        NCCL_IB_HCA: "@fabric.nccl-ib-hca"
+        NCCL_IB_HCA: "@fabric.rdma-devices"
         VLLM_RANDOMIZE_DP_DUMMY_INPUTS: "1"
         VLLM_PREFIX_CACHE_RETENTION_INTERVAL: "32768"
         VLLM_USE_BREAKABLE_CUDAGRAPH: "0"
@@ -139,7 +139,7 @@ base:
         NCCL_CUMEM_ENABLE: "1"
         NCCL_MNNVL_ENABLE: "1"
         NCCL_NVLS_ENABLE: "1"
-        NCCL_IB_HCA: "@fabric.nccl-ib-hca"
+        NCCL_IB_HCA: "@fabric.rdma-devices"
         VLLM_RANDOMIZE_DP_DUMMY_INPUTS: "1"
         VLLM_PREFIX_CACHE_RETENTION_INTERVAL: "32768"
         VLLM_DSV4_MEGA_FP8_COMBINE: "1"

inferencex-e2e/benchmarks/multi_node/srt-slurm-recipes/dsv4/vllm/gb300-fp4/agentx/agg-variants.yaml

@@ -45,7 +45,7 @@ base:
           global_segment_size: "150GB"
           local_buffer_size: "4GB"
           protocol: "rdma"
-          device_name: "@fabric.mooncake-devices"
+          device_name: "@fabric.rdma-devices"
           mode: "embedded"
           enable_offload: false
   frontend:
@@ -91,7 +91,7 @@ base:
         UCX_MEMTYPE_CACHE: "n"
         UCX_NET_DEVICES: "@fabric.ucx-net-devices"
         UCX_TLS: "rc,cuda_copy"
-        NCCL_IB_HCA: "@fabric.nccl-ib-hca"
+        NCCL_IB_HCA: "@fabric.rdma-devices"
         NCCL_P2P_LEVEL: "NVL"
         MC_ENABLE_DEST_DEVICE_AFFINITY: "1"
         MC_STORE_CLIENT_METRIC: "1"

inferencex-e2e/benchmarks/multi_node/srt-slurm-recipes/dsv4/vllm/gb300-fp4/agentx/disagg-1p4d-dep4-tp8-c4-mtp.yaml

@@ -47,7 +47,7 @@ environment:
   NCCL_CUMEM_ENABLE: "1"
   NCCL_MNNVL_ENABLE: "1"
   NCCL_NVLS_ENABLE: "1"
-  NCCL_IB_HCA: "@fabric.nccl-ib-hca"
+  NCCL_IB_HCA: "@fabric.rdma-devices"
 
 slurm:
   time_limit: "08:00:00"

inferencex-e2e/benchmarks/multi_node/srt-slurm-recipes/dsv4/vllm/gb300-fp4/agentx/disagg-mtp-variants.yaml

@@ -42,7 +42,7 @@ base:
           global_segment_size: "180GB"
           local_buffer_size: "4GB"
           protocol: "rdma"
-          device_name: "@fabric.mooncake-devices"
+          device_name: "@fabric.rdma-devices"
           mode: "embedded"
           enable_offload: false
   frontend:
@@ -81,7 +81,7 @@ base:
         NCCL_MNNVL_ENABLE: "1"
         NCCL_NVLS_ENABLE: "1"
         VLLM_RANDOMIZE_DP_DUMMY_INPUTS: "1"
-        NCCL_IB_HCA: "@fabric.nccl-ib-hca"
+        NCCL_IB_HCA: "@fabric.rdma-devices"
         VLLM_USE_BREAKABLE_CUDAGRAPH: "0"
         VLLM_CONNECTOR_PREFETCH_DEPTH: "8"
         VLLM_DSV4_MEGA_FP8_COMBINE: "1"
@@ -137,7 +137,7 @@ base:
         NCCL_MNNVL_ENABLE: "1"
         NCCL_NVLS_ENABLE: "1"
         VLLM_RANDOMIZE_DP_DUMMY_INPUTS: "1"
-        NCCL_IB_HCA: "@fabric.nccl-ib-hca"
+        NCCL_IB_HCA: "@fabric.rdma-devices"
         VLLM_DSV4_MEGA_FP8_COMBINE: "1"
         MC_ENABLE_DEST_DEVICE_AFFINITY: "1"
         MC_STORE_CLIENT_METRIC: "1"