diff --git a/nvflow/lib/gpu_layout.py b/nvflow/lib/gpu_layout.py
index 4afe6ea..2c3dd02 100644
--- a/nvflow/lib/gpu_layout.py
+++ b/nvflow/lib/gpu_layout.py
@@ -66,15 +66,24 @@ def resolve_gpu_layout(
         Resolved :class:`GpuLayout`.
 
     Raises:
-        ValueError: If ``total_gpus`` is not evenly divisible by
-            ``gpus_per_node``.
+        ValueError: If ``gpus_per_node`` is not a positive integer, if
+            ``total_gpus`` is not a positive integer evenly divisible by
+            ``gpus_per_node``, or if the legacy ``num_nodes``/``num_gpus``
+            fields are not positive integers.
     """
     gpus_per_node = _DEFAULT_GPUS_PER_NODE
     if cluster_config:
         gpus_per_node = cluster_config.get("gpus_per_node", _DEFAULT_GPUS_PER_NODE)
+    if gpus_per_node <= 0:
+        raise ValueError(
+            f"gpus_per_node ({gpus_per_node}) must be a positive integer.  "
+            f"Check the gpus_per_node value in the cluster config."
+        )
 
     if "total_gpus" in config:
         total = config["total_gpus"]
+        if total <= 0:
+            raise ValueError(f"total_gpus ({total}) must be a positive integer.")
         if total % gpus_per_node != 0:
             raise ValueError(
                 f"total_gpus ({total}) is not evenly divisible by "
@@ -86,7 +95,11 @@ def resolve_gpu_layout(
 
     if "num_nodes" in config:
         num_nodes = config["num_nodes"]
+        if num_nodes <= 0:
+            raise ValueError(f"num_nodes ({num_nodes}) must be a positive integer.")
         num_gpus = config.get("num_gpus", gpus_per_node)
+        if num_gpus <= 0:
+            raise ValueError(f"num_gpus ({num_gpus}) must be a positive integer.")
         LOG.debug(
             "Using legacy num_nodes=%d / num_gpus=%d (not portable across clusters)",
             num_nodes,
diff --git a/tests/test_gpu_layout.py b/tests/test_gpu_layout.py
new file mode 100644
index 0000000..c2bf189
--- /dev/null
+++ b/tests/test_gpu_layout.py
@@ -0,0 +1,87 @@
+# Copyright (c) 2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+"""Tests for ``nvflow.lib.gpu_layout``."""
+
+from __future__ import annotations
+
+import pytest
+
+from nvflow.lib.gpu_layout import GpuLayout, resolve_gpu_layout
+
+
+class TestGpuLayout:
+    def test_total_gpus_property(self) -> None:
+        layout = GpuLayout(num_nodes=2, gpus_per_node=8)
+        assert layout.total_gpus == 16
+
+
+class TestResolveGpuLayoutTotalGpus:
+    def test_divides_evenly_with_cluster_config(self) -> None:
+        layout = resolve_gpu_layout({"total_gpus": 16}, {"gpus_per_node": 4})
+        assert layout == GpuLayout(num_nodes=4, gpus_per_node=4)
+
+    def test_uses_default_gpus_per_node_without_cluster_config(self) -> None:
+        layout = resolve_gpu_layout({"total_gpus": 16})
+        assert layout == GpuLayout(num_nodes=2, gpus_per_node=8)
+
+    def test_not_evenly_divisible_raises(self) -> None:
+        with pytest.raises(ValueError, match="not evenly divisible"):
+            resolve_gpu_layout({"total_gpus": 10}, {"gpus_per_node": 8})
+
+    def test_zero_total_gpus_raises(self) -> None:
+        with pytest.raises(ValueError, match="must be a positive integer"):
+            resolve_gpu_layout({"total_gpus": 0}, {"gpus_per_node": 8})
+
+    def test_negative_total_gpus_raises(self) -> None:
+        with pytest.raises(ValueError, match="must be a positive integer"):
+            resolve_gpu_layout({"total_gpus": -8}, {"gpus_per_node": 8})
+
+
+class TestResolveGpuLayoutGpusPerNode:
+    def test_zero_gpus_per_node_raises_value_error_not_zero_division(self) -> None:
+        with pytest.raises(ValueError, match="gpus_per_node.*must be a positive integer"):
+            resolve_gpu_layout({"total_gpus": 16}, {"gpus_per_node": 0})
+
+    def test_negative_gpus_per_node_raises(self) -> None:
+        with pytest.raises(ValueError, match="gpus_per_node.*must be a positive integer"):
+            resolve_gpu_layout({"total_gpus": 16}, {"gpus_per_node": -4})
+
+
+class TestResolveGpuLayoutLegacyNumNodes:
+    def test_num_nodes_with_explicit_num_gpus(self) -> None:
+        layout = resolve_gpu_layout({"num_nodes": 3, "num_gpus": 4})
+        assert layout == GpuLayout(num_nodes=3, gpus_per_node=4)
+
+    def test_num_nodes_without_num_gpus_falls_back_to_gpus_per_node(self) -> None:
+        layout = resolve_gpu_layout({"num_nodes": 3}, {"gpus_per_node": 2})
+        assert layout == GpuLayout(num_nodes=3, gpus_per_node=2)
+
+    def test_zero_num_nodes_raises(self) -> None:
+        with pytest.raises(ValueError, match="num_nodes.*must be a positive integer"):
+            resolve_gpu_layout({"num_nodes": 0})
+
+    def test_zero_num_gpus_raises(self) -> None:
+        with pytest.raises(ValueError, match="num_gpus.*must be a positive integer"):
+            resolve_gpu_layout({"num_nodes": 2, "num_gpus": 0})
+
+
+class TestResolveGpuLayoutDefault:
+    def test_neither_field_defaults_to_single_node(self) -> None:
+        layout = resolve_gpu_layout({})
+        assert layout == GpuLayout(num_nodes=1, gpus_per_node=8)
+
+    def test_neither_field_uses_cluster_gpus_per_node(self) -> None:
+        layout = resolve_gpu_layout({}, {"gpus_per_node": 4})
+        assert layout == GpuLayout(num_nodes=1, gpus_per_node=4)