Re-configure benchmarking devices & add markers to `bench_cugraph_uni…

…form_neighbor_sample` (#4561)

This PR re-enables the `benchmarks/cugraph/pytest-based/bench_cugraph_uniform_neighbor_sample.py` benchmark in the MNMG nightlies. 

This benchmark was previously being skipped due to the fact that the benchmarks are configured to run with Pytest markers `"managedmem_off and poolallocator_on"`. Thanks to @jameslamb for spotting this inside `conftest.py`

By adding the missing markers to the test (and removing some `dgx` machine specific dask-client configs), the benchmark should properly run in the nightly jobs.

Authors:
  - Ralph Liu (https://github.com/nv-rliu)

Approvers:
  - Rick Ratzel (https://github.com/rlratzel)
  - James Lamb (https://github.com/jameslamb)

URL: #4561

benchmarks/cugraph/pytest-based/bench_cugraph_uniform_neighbor_sample.py

@@ -48,6 +48,11 @@
 _seed = 42
 
 
+###############################################################################
+# Helpers
+###############################################################################
+
+
 def create_graph(graph_data):
     """
     Create a graph instance based on the data to be loaded/generated, return a
@@ -107,17 +112,13 @@ def create_mg_graph(graph_data):
     Create a graph instance based on the data to be loaded/generated, return a
     tuple containing (graph_obj, num_verts, client, cluster)
     """
-    # range starts at 1 to let let 0 be used by benchmark/client process
-    visible_devices = os.getenv("DASK_WORKER_DEVICES", "1,2,3,4")
-
     (client, cluster) = start_dask_client(
         # enable_tcp_over_ucx=True,
         # enable_infiniband=False,
         # enable_nvlink=True,
         # enable_rdmacm=False,
         protocol="ucx",
         rmm_pool_size="28GB",
-        dask_worker_devices=visible_devices,
     )
     rmm.reinitialize(pool_allocator=True)
 
@@ -261,6 +262,9 @@ def uns_func(*args, **kwargs):
 
 ################################################################################
 # Benchmarks
+###############################################################################
+@pytest.mark.managedmem_off
+@pytest.mark.poolallocator_on
 @pytest.mark.parametrize("batch_size", params.batch_sizes.values())
 @pytest.mark.parametrize("fanout", [params.fanout_10_25, params.fanout_5_10_15])
 @pytest.mark.parametrize(