ci: add pytorch 2.12 and 2.13 to cpu tests (#21835)

* ci: add pytorch 2.12 and 2.13 to cpu tests

* test: avoid global privateuse1 backend pollution breaking CI on PyTorch 2.12/2.13

* test: simplify with monkeypatch.setitem

Author: bhimrazy

.github/workflows/ci-tests-fabric.yml

@@ -40,24 +40,29 @@ jobs:
       matrix:
         os: [macOS-14, ubuntu-22.04, windows-2022]
         config:
+          # TODO: follow-up - prune the oldest PyTorch minors (2.1-2.5) now that we test up to 2.13
           # Test unified "lightning" package with PyTorch 2.1-2.5
           - { pkg-name: "lightning", python-version: "3.10", pytorch-version: "2.1" }
           - { pkg-name: "lightning", python-version: "3.11", pytorch-version: "2.2.2" }
           - { pkg-name: "lightning", python-version: "3.11", pytorch-version: "2.3" }
           - { pkg-name: "lightning", python-version: "3.12.7", pytorch-version: "2.4.1" }
           - { pkg-name: "lightning", python-version: "3.12.7", pytorch-version: "2.5.1" }
 
-          # Test "fabric" package with PyTorch 2.6-2.11
+          # Test "fabric" package with PyTorch 2.6-2.13
           - { pkg-name: "fabric", python-version: "3.12.7", pytorch-version: "2.6" }
           - { pkg-name: "fabric", python-version: "3.12", pytorch-version: "2.7" }
           - { pkg-name: "fabric", python-version: "3.12", pytorch-version: "2.8" }
           - { pkg-name: "fabric", python-version: "3.12", pytorch-version: "2.9" }
           - { pkg-name: "fabric", python-version: "3.12", pytorch-version: "2.10" }
           - { pkg-name: "fabric", python-version: "3.12", pytorch-version: "2.11" }
+          - { pkg-name: "fabric", python-version: "3.12", pytorch-version: "2.12" }
+          - { pkg-name: "fabric", python-version: "3.12", pytorch-version: "2.13" }
           - { pkg-name: "fabric", python-version: "3.13", pytorch-version: "2.8" }
           - { pkg-name: "fabric", python-version: "3.13", pytorch-version: "2.9" }
           - { pkg-name: "fabric", python-version: "3.13", pytorch-version: "2.10" }
           - { pkg-name: "fabric", python-version: "3.13", pytorch-version: "2.11" }
+          - { pkg-name: "fabric", python-version: "3.13", pytorch-version: "2.12" }
+          - { pkg-name: "fabric", python-version: "3.13", pytorch-version: "2.13" }
 
           # Test minimum supported versions (oldest)
           - { pkg-name: "fabric", pytorch-version: "2.1", requires: "oldest" }

.github/workflows/ci-tests-pytorch.yml

@@ -44,24 +44,29 @@ jobs:
       matrix:
         os: [macOS-14, ubuntu-22.04, windows-2022]
         config:
+          # TODO: follow-up - prune the oldest PyTorch minors (2.1-2.5) now that we test up to 2.13
           # Test unified "lightning" package with PyTorch 2.1-2.5
           - { pkg-name: "lightning", python-version: "3.10", pytorch-version: "2.1" }
           - { pkg-name: "lightning", python-version: "3.11", pytorch-version: "2.2.2" }
           - { pkg-name: "lightning", python-version: "3.11", pytorch-version: "2.3" }
           - { pkg-name: "lightning", python-version: "3.12.7", pytorch-version: "2.4.1" }
           - { pkg-name: "lightning", python-version: "3.12.7", pytorch-version: "2.5.1" }
 
-          # Test "pytorch" package with PyTorch 2.6-2.11
+          # Test "pytorch" package with PyTorch 2.6-2.13
           - { pkg-name: "pytorch", python-version: "3.12.7", pytorch-version: "2.6" }
           - { pkg-name: "pytorch", python-version: "3.12", pytorch-version: "2.7" }
           - { pkg-name: "pytorch", python-version: "3.12", pytorch-version: "2.8" }
           - { pkg-name: "pytorch", python-version: "3.12", pytorch-version: "2.9" }
           - { pkg-name: "pytorch", python-version: "3.12", pytorch-version: "2.10" }
           - { pkg-name: "pytorch", python-version: "3.12", pytorch-version: "2.11" }
+          - { pkg-name: "pytorch", python-version: "3.12", pytorch-version: "2.12" }
+          - { pkg-name: "pytorch", python-version: "3.12", pytorch-version: "2.13" }
           - { pkg-name: "pytorch", python-version: "3.13", pytorch-version: "2.8" }
           - { pkg-name: "pytorch", python-version: "3.13", pytorch-version: "2.9" }
           - { pkg-name: "pytorch", python-version: "3.13", pytorch-version: "2.10" }
           - { pkg-name: "pytorch", python-version: "3.13", pytorch-version: "2.11" }
+          - { pkg-name: "pytorch", python-version: "3.13", pytorch-version: "2.12" }
+          - { pkg-name: "pytorch", python-version: "3.13", pytorch-version: "2.13" }
 
           # Test minimum supported versions (oldest)
           - { pkg-name: "pytorch", pytorch-version: "2.1", requires: "oldest" }

tests/tests_fabric/utilities/test_distributed.py

@@ -245,22 +245,14 @@ def test_init_dist_connection_registers_destruction_handler(_, atexit_mock):
     atexit_mock.register.assert_not_called()
 
 
-def test_get_default_process_group_backend_for_device():
+def test_get_default_process_group_backend_for_device(monkeypatch):
     """Test that each device type maps to its correct default process group backend."""
-    # register a custom backend for test
-    torch.utils.rename_privateuse1_backend("pcu")
-
-    def mock_backend(store, group_rank, group_size, timeout):
-        pass
-
-    torch.distributed.Backend.register_backend(
-        "pccl",
-        lambda store, group_rank, group_size, timeout: mock_backend(store, group_rank, group_size, timeout),
-        devices=["pcu"],
-    )
+    # patch the map instead of registering a real backend: `rename_privateuse1_backend` is irreversible
+    # and would leak a dangling "pcu" device into subsequent tests
+    monkeypatch.setitem(torch.distributed.Backend.default_device_backend_map, "pcu", "pccl")
 
     # test that the default backend is correctly set for each device
-    devices = [torch.device("cpu"), torch.device("cuda:0"), torch.device("pcu:0")]
+    devices = [torch.device("cpu"), torch.device("cuda:0"), Mock(type="pcu")]
     backends = ["gloo", "nccl", "pccl"]
     for device, backend in zip(devices, backends):
         assert _get_default_process_group_backend_for_device(device) == backend