[ROCm] [CI] Add new fusion test cases that are relevant to vLLM IR Ops (#34307)

Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
Signed-off-by: vllmellm <vllm.ellm@embeddedllm.com>
Co-authored-by: vllmellm <vllm.ellm@embeddedllm.com>

tests/compile/fusions_e2e/test_tp2_async_tp.py

@@ -5,6 +5,7 @@ from collections.abc import Callable
 import pytest
 
 from vllm.config import PassConfig
+from vllm.platforms import current_platform
 
 from ...utils import multi_gpu_test
 from .common import (
@@ -23,6 +24,8 @@ from .models import (
     qwen3_a3b,
 )
 
+pytestmark = pytest.mark.skipif(not current_platform.is_cuda(), reason="Only test CUDA")
+
 
 @multi_gpu_test(num_gpus=2)
 @pytest.mark.parametrize(