[ROCm][CI] Fix tool use test stability - disable skinny GEMM, prefix caching, eliminate batch variance

docs/design/attention_backends.md

@@ -171,7 +171,7 @@ Priority is **1 = highest** (tried first).
 | `FLASH_ATTN` | FA4* | fp16, bf16 | `auto`, `bfloat16` | %16 | Any | ❌ | ❌ | ✅ | All | ≥10.0 |
 | `FLASH_ATTN_DIFFKV` |  | fp16, bf16 | `auto` | Any | Any | ❌ | ❌ | ✅ | Decoder | Any |
 | `FLEX_ATTENTION` |  | fp16, bf16, fp32 | `auto`, `bfloat16` | Any | Any | ❌ | ✅ | ❌ | Decoder, Encoder Only | Any |
-| `ROCM_AITER_FA` |  | fp16, bf16 | `auto`, `bfloat16`, `fp8`, `fp8_e4m3`, `fp8_e5m2` | 16, 32 | 64, 128, 256 | ❌ | ❌ | ❌ | Decoder | N/A |
+| `ROCM_AITER_FA` |  | fp16, bf16 | `auto`, `bfloat16`, `fp8`, `fp8_e4m3`, `fp8_e5m2` | 16, 32 | 64, 128, 256 | ❌ | ❌ | ❌ | Decoder, Enc-Dec | N/A |
 | `ROCM_AITER_UNIFIED_ATTN` |  | fp16, bf16 | `auto` | %16 | Any | ✅ | ✅ | ❌ | All | N/A |
 | `ROCM_ATTN` |  | fp16, bf16, fp32 | `auto`, `bfloat16`, `fp8`, `fp8_e4m3`, `fp8_e5m2` | 16, 32, 544 | 32, 64, 80, 96, 128, 160, 192, 224, 256 | ✅ | ✅ | ❌ | All | N/A |
 | `TREE_ATTN` |  | fp16, bf16 | `auto` | %16 | 32, 64, 96, 128, 160, 192, 224, 256 | ❌ | ❌ | ❌ | Decoder | Any |

requirements/rocm-test.txt

@@ -108,3 +108,5 @@ bitsandbytes==0.49.2
 tensorizer==2.10.1
 # Multi-modal models test (`allendou/FireRedASR2-LLM-vllm`)
 kaldi-native-fbank==1.22.3
+# Pinning numpy version
+numpy==2.2.6

tests/entrypoints/openai/test_completion_with_function_calling.py

@@ -9,14 +9,13 @@
 import pytest
 import pytest_asyncio
 
-from vllm.platforms import current_platform
-
 # downloading lora to test lora requests
-from ...utils import RemoteOpenAIServer
+from ...utils import ROCM_ENV_OVERRIDES, ROCM_EXTRA_ARGS, RemoteOpenAIServer
 
 # any model with a chat template should work here
 MODEL_NAME = "Qwen/Qwen3-0.6B"
 
+
 tools = [
     {
         "type": "function",
@@ -142,19 +141,11 @@ def server():
         "--gpu-memory-utilization",
         "0.4",
         "--enforce-eager",
-    ]
+    ] + ROCM_EXTRA_ARGS
 
-    rocm_args = {
-        "--max-num-seqs": "1",
-        "--no-enable-prefix-caching": None,
-    }
-    if current_platform.is_rocm():
-        for k, v in rocm_args.items():
-            args.append(k)
-            if v is not None:
-                args.append(v)
-
-    with RemoteOpenAIServer(MODEL_NAME, args) as remote_server:
+    with RemoteOpenAIServer(
+        MODEL_NAME, args, env_dict=ROCM_ENV_OVERRIDES
+    ) as remote_server:
         yield remote_server
 
 
@@ -239,12 +230,13 @@ def k2_server():
         "qwen3",
         "--gpu-memory-utilization",
         "0.4",
-    ]
+    ] + ROCM_EXTRA_ARGS
     # hack to test kimi_k2 tool use tool_id format.
     # avoid error in is_deepseek_mla check by setting kv_lora_rank=null
     with RemoteOpenAIServer(
         MODEL_NAME,
         args,
+        env_dict=ROCM_ENV_OVERRIDES,
         override_hf_configs={"model_type": "kimi_k2", "kv_lora_rank": None},
     ) as remote_server:
         yield remote_server

tests/utils.py

@@ -109,6 +109,20 @@ def _nvml():
 VLLM_PATH = Path(__file__).parent.parent
 """Path to root of the vLLM repository."""
 
+# ROCm: disable skinny GEMM to avoid non-deterministic results from
+# atomic reductions in wvSplitKrc kernel.
+# See: https://github.com/vllm-project/vllm/pull/33493#issuecomment-3906083975
+ROCM_ENV_OVERRIDES = (
+    {"VLLM_ROCM_USE_SKINNY_GEMM": "0"} if current_platform.is_rocm() else {}
+)
+# ROCm: disable prefix caching and eliminate batch variance to reduce
+# test flakiness.
+ROCM_EXTRA_ARGS = (
+    ["--no-enable-prefix-caching", "--max-num-seqs", "1"]
+    if current_platform.is_rocm()
+    else []
+)
+
 
 class RemoteVLLMServer:
     """Base class for launching vLLM server subprocesses for testing.

vllm/v1/attention/backends/rocm_aiter_fa.py

@@ -741,6 +741,14 @@ class AiterFlashAttentionBackend(AttentionBackend):
         "fp8_e5m2",
     ]
 
+    @classmethod
+    def supports_attn_type(cls, attn_type: str) -> bool:
+        """ROCM AITER FA supports decoder and encoder-decoder (cross) attention."""
+        return attn_type in (
+            AttentionType.DECODER,
+            AttentionType.ENCODER_DECODER,
+        )
+
     @staticmethod
     def get_supported_kernel_block_sizes() -> list[int | MultipleOf]:
         return [16, 32]