test(nemotron): align fusion checks with job configs

Signed-off-by: Malay Nagda <malayn@nvidia.com>

Author: malay-nagda

tests/unit_tests/recipes/test_nemotron_3_5_lightning_perf_recipes.py

@@ -197,10 +197,10 @@ def test_gb200_mxfp8_enables_cutedsl_fusion() -> None:
     assert cfg.model.use_transformer_engine_op_fuser is True
     assert cfg.model.moe_mlp_glu_interleave_size == 32
     assert cfg.model.high_priority_a2a_comm_stream is False
-    assert cfg.model.moe_hybridep_num_sms_preprocessing != 32
+    assert cfg.model.moe_hybridep_num_sms_preprocessing == 108
     assert cfg.mixed_precision.fp8_dot_product_attention is True
-    assert cfg.comm_overlap.overlap_moe_expert_parallel_comm is not True
-    assert cfg.comm_overlap.delay_wgrad_compute is not True
+    assert cfg.comm_overlap.overlap_moe_expert_parallel_comm is False
+    assert cfg.comm_overlap.delay_wgrad_compute is False
 
 
 def test_gb200_mxfp8_fsdp_skips_cutedsl_fusion() -> None:

tests/unit_tests/recipes/test_nemotronh_recipes.py

@@ -54,6 +54,8 @@ class _FakeModelProvider:
 
     def __init__(self) -> None:
         self.vocab_size = 256
+        self.high_priority_a2a_comm_stream = False
+        self.moe_hybridep_num_sms_preprocessing = 108
 
     def finalize(self) -> None:
         return None
@@ -163,32 +165,40 @@ def test_nemotron_3_nano_gb200_defers_vocab_size_to_training_tokenizer():
 
 
 @pytest.mark.parametrize(
-    ("module_name", "factory_name"),
+    ("module_name", "factory_name", "has_comm_overlap"),
     [
-        (
+        pytest.param(
             "megatron.bridge.perf_recipes.nemotronh.gb200.nemotronh",
             "nemotron_3_nano_pretrain_8gpu_gb200_fp8mx_config",
+            True,
+            id="nano-gb200",
         ),
-        (
+        pytest.param(
             "megatron.bridge.perf_recipes.nemotronh.gb300.nemotronh",
             "nemotron_3_nano_pretrain_8gpu_gb300_fp8mx_config",
+            True,
+            id="nano-gb300",
         ),
-        (
+        pytest.param(
             "megatron.bridge.perf_recipes.nemotronh.gb200.nemotronh",
             "nemotron_3_super_pretrain_64gpu_gb200_fp8mx_config",
+            False,
+            id="super-gb200",
         ),
-        (
+        pytest.param(
             "megatron.bridge.perf_recipes.nemotronh.gb300.nemotronh",
             "nemotron_3_super_pretrain_64gpu_gb300_fp8mx_config",
+            False,
+            id="super-gb300",
         ),
     ],
-    ids=lambda recipe: recipe.rsplit(".", 1)[-1],
 )
 def test_nemotron_3_gb_mxfp8_perf_recipes_enable_cutedsl_fusion(
     module_name: str,
     factory_name: str,
+    has_comm_overlap: bool,
 ) -> None:
-    """GB200 and GB300 MXFP8 recipes enable CutDSL without MoE A2A overlap."""
+    """GB MXFP8 recipes match the measured CutDSL and MoE overlap settings."""
     module = importlib.import_module(module_name)
     cfg = getattr(module, factory_name)()
 
@@ -197,10 +207,14 @@ def test_nemotron_3_gb_mxfp8_perf_recipes_enable_cutedsl_fusion(
     assert cfg.model.use_transformer_engine_op_fuser is True
     assert cfg.model.moe_mlp_glu_interleave_size == 32
     assert cfg.model.high_priority_a2a_comm_stream is False
-    assert cfg.model.moe_hybridep_num_sms_preprocessing != 32
+    assert cfg.model.moe_hybridep_num_sms_preprocessing == 108
     assert cfg.mixed_precision.fp8_dot_product_attention is True
-    assert cfg.comm_overlap.overlap_moe_expert_parallel_comm is not True
-    assert cfg.comm_overlap.delay_wgrad_compute is not True
+    if has_comm_overlap:
+        assert cfg.comm_overlap is not None
+        assert cfg.comm_overlap.overlap_moe_expert_parallel_comm is False
+        assert cfg.comm_overlap.delay_wgrad_compute is False
+    else:
+        assert cfg.comm_overlap is None
 
 
 def test_nemotron_3_super_64gpu_gb200_matches_benchmark_hardware_configuration():