[PyTorch] Check if FP8 block scaling is supported in tests (#1727)

* Check if FP8 block scaling is supported in tests
Signed-off-by: Tim Moon <tmoon@nvidia.com>

* Update tests/pytorch/distributed/test_cast_master_weights_to_fp8.py
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

---------
Signed-off-by: Tim Moon <tmoon@nvidia.com>
Signed-off-by: Tim Moon <4406448+timmoon10@users.noreply.github.com>

tests/pytorch/distributed/test_cast_master_weights_to_fp8.py

@@ -15,6 +15,9 @@ if torch.cuda.device_count() < 2:
     pytest.skip("cast_master_weights_to_fp8 test needs at least 2 GPUs.")
 
 fp8_available, reason_for_no_fp8 = FP8GlobalStateManager.is_fp8_available()
+fp8_block_scaling_available, reason_for_no_fp8_block_scaling = (
+    FP8GlobalStateManager.is_fp8_block_scaling_available()
+)
 
 TEST_ROOT = Path(__file__).parent.resolve()
 NUM_PROCS: int = min(2, torch.cuda.device_count())
@@ -30,6 +33,8 @@ def _run_test(quantization):
 
 @pytest.mark.parametrize("quantization", ["fp8", "fp8_cs", "fp8_block"])
 def test_cast_master_weights_to_fp8(quantization):
-    if not fp8_available:
+    if quantization in ("fp8", "fp8_cs") and not fp8_available:
         pytest.skip(reason_for_no_fp8)
+    if quantization == "fp8_block" and not fp8_block_scaling_available:
+        pytest.skip(reason_for_no_fp8_block_scaling)
     _run_test(quantization)