[CI][ROCm] Add a monokernels test area with a group per AMD monokernel

.buildkite/test_areas/kernels.yaml

@@ -1076,27 +1076,3 @@ steps:
   - amd-smi
   - python3 examples/basic/offline_inference/chat.py --attention-backend TRITON_ATTN
   - pytest -v -s tests/kernels/attention/test_attention_selector.py
-
-- label: ":amd: (MI300) DeepSeek-V4.1 gfx942 Mono Decode Kernels"
-  key: dsv41-mono-decode-gfx942-amd
-  dind: false
-  device: mi300_1
-  num_devices: 1
-  timeout_in_minutes: 30
-  working_dir: /vllm-workspace/
-  depends_on:
-  - image-build-amd
-  source_file_dependencies:
-  - vllm/models/deepseek_v41/amd/
-  - vllm/model_executor/layers/sparse_attn_indexer.py
-  - vllm/v1/attention/ops/rocm_aiter_mla_sparse.py
-  - vllm/v1/worker/gpu/spec_decode/dspark/
-  - tests/kernels/models/amd/test_dsv41_gfx942_topk.py
-  - tests/kernels/models/amd/test_dsv41_gfx942_cand_logits.py
-  - tests/kernels/models/amd/test_dsv41_gfx942_sparse_indexer.py
-  - tests/kernels/models/amd/test_dsv41_gfx942_decode_meta.py
-  - tests/kernels/models/amd/test_dsv41_gfx942_fused_markov.py
-  - tests/kernels/models/amd/test_dsv41_gfx942_draft_lm_head.py
-  - tests/kernels/models/amd/test_dsv41_mono_gfx942.py
-  commands:
-  - pytest -v -s tests/kernels/models/amd/test_dsv41_gfx942_topk.py tests/kernels/models/amd/test_dsv41_gfx942_cand_logits.py tests/kernels/models/amd/test_dsv41_gfx942_sparse_indexer.py tests/kernels/models/amd/test_dsv41_gfx942_decode_meta.py tests/kernels/models/amd/test_dsv41_gfx942_fused_markov.py tests/kernels/models/amd/test_dsv41_gfx942_draft_lm_head.py tests/kernels/models/amd/test_dsv41_mono_gfx942.py

.buildkite/test_areas/models_basic.yaml

@@ -103,36 +103,18 @@ steps:
       - tests/models/test_deepseek_v4_dspark_rocm.py
       - tests/models/test_deepseek_v4_rocm_wo_a.py
       - tests/models/test_deepseek_v4_rocm_compressor_gemm_fusion.py
-      - tests/models/test_dsv41_mono_decode.py
+      - tests/models/minimax_m3/test_index_cache_overrides.py
       - tests/models/test_muse_spark_rocm_loading.py
       - tests/models/test_deepseek_v4_vl_stream_weights.py
       - tests/models/hy_v4/test_triton_ihc.py
       - vllm/models/muse_spark/
       - tests/models/test_muse_spark_logits_local.py
       - tests/models/test_muse_spark_rocm_shared_rope.py
       commands:
-      - pytest -v -s models/test_terratorch.py models/transformers/test_backend.py models/test_registry.py models/test_deepseek_v4_vl_rocm.py models/transformers/test_layer_registry_e2e.py models/test_deepseek_v41_replay_start.py models/test_deepseek_v41_decoder_replay_layers.py models/test_deepseek_v41_replay_batch.py models/test_deepseek_v41_amd_replay_forward.py models/test_deepseek_v4_dspark_rocm.py models/test_deepseek_v4_rocm_wo_a.py models/test_deepseek_v4_rocm_compressor_gemm_fusion.py models/test_dsv41_mono_decode.py models/test_muse_spark_rocm_loading.py models/test_deepseek_v4_vl_stream_weights.py models/test_muse_spark_logits_local.py models/test_muse_spark_rocm_shared_rope.py
+      - pytest -v -s models/test_terratorch.py models/transformers/test_backend.py models/test_registry.py models/test_deepseek_v4_vl_rocm.py models/transformers/test_layer_registry_e2e.py models/test_deepseek_v41_replay_start.py models/test_deepseek_v41_decoder_replay_layers.py models/test_deepseek_v41_replay_batch.py models/test_deepseek_v41_amd_replay_forward.py models/test_deepseek_v4_dspark_rocm.py models/test_deepseek_v4_rocm_wo_a.py models/test_deepseek_v4_rocm_compressor_gemm_fusion.py models/minimax_m3/test_index_cache_overrides.py models/test_muse_spark_rocm_loading.py models/test_deepseek_v4_vl_stream_weights.py models/test_muse_spark_logits_local.py models/test_muse_spark_rocm_shared_rope.py
       - VLLM_ROCM_USE_AITER=1 pytest -v -s models/test_hyv4_rocm.py -m 'not distributed'
       - pytest -v -s models/hy_v4/test_triton_ihc.py
 
-- label: ":amd: (MI355) MiniMax M3 MonoKernel"
-  key: minimax-m3-monokernel-mi355
-  dind: false
-  timeout_in_minutes: 45
-  device: mi355_4
-  num_devices: 4
-  working_dir: "/vllm-workspace/tests"
-  depends_on:
-  - image-build-amd
-  source_file_dependencies:
-  - vllm/models/minimax_m3/amd/
-  - tests/models/minimax_m3/amd_specialized_decode/
-  - tests/models/minimax_m3/test_index_cache_overrides.py
-  - vllm/model_executor/models/config.py
-  commands:
-  - pytest -v -s models/minimax_m3/amd_specialized_decode
-  - pytest -v -s models/minimax_m3/test_index_cache_overrides.py
-
 - label: ":nvidia: (B200) Inkling"
   key: inkling-unit-tests-b200
   timeout_in_minutes: 40
@@ -202,7 +184,7 @@ steps:
       # The ROCm KDA/AttnRes kernels are gated on gfx950, so this has to be MI355.
       # The kernels/ file the NVIDIA job adds is CUDA-only (it imports
       # cuda.bindings.driver), so it is deliberately left out here.
-      - pytest -v -s models/kimi_k3
+      - pytest -v -s models/kimi_k3 --ignore=models/kimi_k3/amd_specialized_decode
 
 - label: ":nvidia: (4xB200) [Nightly] Kimi K3 Prefix Cache"
   key: kimi-k3-prefix-cache-4xb200
@@ -328,44 +310,6 @@ steps:
   commands:
     - pytest -v -s models/test_utils.py models/test_vision.py models/test_adapters.py models/test_qwen3_5_mtp_config.py models/transformers/fusers/ models/transformers/test_dynamic_rope.py model_executor/layers/test_activation.py models/test_language_model_cache_is_weak.py models/deepseek_v32/test_sequence_parallel.py models/test_deepseek_v4_fi_moe_ep.py models/test_dspark_mla.py models/transformers/test_layer_registry.py models/test_alice_ai.py models/test_alice_ai_config.py models/test_beam_weight_mapping.py models/test_beam_numerics.py models/test_cohere2_moe_parallel_block.py
 
-- label: ':amd: (MI355) DeepSeek-V4.1 Mono Decode Numerics'
-  key: dsv41-mono-decode-numerics-amd
-  dind: false
-  device: mi355_4
-  num_devices: 4
-  timeout_in_minutes: 40
-  working_dir: /vllm-workspace/
-  depends_on:
-  - image-build-amd
-  source_file_dependencies:
-  - docker/Dockerfile.rocm_base
-  - vllm/models/deepseek_v41/amd/
-  - vllm/models/deepseek_v41/nvidia/model_state.py
-  - tests/models/test_dsv41_mono_numerics.py
-  - tests/models/test_dsv41_mono_layers.py
-  commands:
-  - pytest -v -s tests/models/test_dsv41_mono_numerics.py
-  - pytest -v -s tests/models/test_dsv41_mono_layers.py
-
-- label: ':amd: (MI300) DeepSeek-V4.1 gfx942 Mono Decode Layers'
-  key: dsv41-mono-decode-layers-gfx942-amd
-  dind: false
-  device: mi300_4
-  num_devices: 4
-  timeout_in_minutes: 40
-  working_dir: /vllm-workspace/
-  depends_on:
-  - image-build-amd
-  source_file_dependencies:
-  - docker/Dockerfile.rocm_base
-  - vllm/models/deepseek_v41/amd/
-  - vllm/v1/attention/ops/rocm_aiter_mla_sparse.py
-  - tests/models/test_dsv41_mono_numerics.py
-  - tests/models/test_dsv41_mono_layers.py
-  commands:
-  - pytest -v -s tests/models/test_dsv41_mono_numerics.py
-  - pytest -v -s tests/models/test_dsv41_mono_layers.py
-
 - label: ":nvidia: (4xB200) DeepSeek-V4.1 Specialized Decode"
   key: dsv41-nvidia-specialized-decode-4xb200
   # Nightly only: 4-GPU B200 pods wait hours behind smaller jobs on per-PR runs.
@@ -390,18 +334,3 @@ steps:
   - pytest -v -s models/deepseek_v41/nvidia/specialized_decode/test_kernels.py
   - pytest -v -s models/deepseek_v41/nvidia/specialized_decode/test_numerics.py
   - pytest -v -s models/deepseek_v41/nvidia/specialized_decode/test_layers.py
-
-- label: ':amd: (MI355) Kimi-K3 Specialized Decode'
-  key: kimi-k3-specialized-decode-amd
-  dind: false
-  device: mi355_8
-  num_devices: 8
-  timeout_in_minutes: 60
-  working_dir: /vllm-workspace/
-  depends_on:
-  - image-build-amd
-  source_file_dependencies:
-  - vllm/models/kimi_k3/amd/
-  - tests/models/kimi_k3/amd_specialized_decode/
-  commands:
-  - pytest -v -s tests/models/kimi_k3/amd_specialized_decode