[CI][Refactor] Merge test CI workflow files into one (#973)

* refactor: merge test CI workflow files into one

* chore: set `UV_INDEX_STRATEGY=unsafe-best-match`

* feat: add AST test with Python 3.8

* feat: implement manual caching mechanism for self-hosted runners

* refactor: simplify cache logic for self-hosted runners

* chore: clear uv cache on failure

* chore: print format.sh output to logs

* chore: improve uv caching

* chore: disable parallel test

* chore: use `PYTHONDEVMODE=1` in CI

* feat: enable coredump generation

* fix: fix perfbench condition

* Revert "feat: enable coredump generation"

This reverts commit c52da65cb572932e09905d08c43a39ec3cf47c54.

* chore: move example CI down

* Revert "chore: move example CI down"

This reverts commit 9d8e65055e01d955c5268a9a6705d270c2de0d57.

* chore: skip example `test_example_mha_sink_bwd_bhsd`

* chore: skip example `test_example_gqa_sink_bwd_bhsd`

* fix: fix example argument passing

* fix: loosen test criteria

* chore: rename `CMAKE_CONFIG...

requirements-test-rocm.txt

+# Lint requirements
+--requirement requirements-lint.txt
+
+# Common test requirements
+--requirement requirements-test.txt
+
+# ROCm specific requirements
+# Currently: none

requirements-test.txt

-# lint requirements
--r requirements-lint.txt
-# build requirements
-Cython
+# Lint requirements
+--requirement requirements-lint.txt
+
+# Build requirements
+cython
 cmake>=3.26
-# runtime requirements
+setuptools
+scikit-build-core
+ninja
+
+# Runtime requirements
+--requirement requirements.txt
+
+# Test requirements
 cffi
 cpplint
-Cython
+cython
 docutils
 dtlib
-numpy>=1.23.5
-pytest>=6.2.4
-pytest_xdist>=2.2.1
+einops
+packaging>=21.0
+pytest-xdist>=2.2.1
 pytest-durations
 pytest-timeout
-packaging>=21.0
-PyYAML
-tqdm>=4.62.3
-typing_extensions>=4.10.0
+pytest>=6.2.4
+pyyaml
 requests
-cloudpickle
-ml_dtypes
-psutil
-torch
-tabulate
-wheel
-setuptools
-einops
 scipy
+tabulate
 tornado
+wheel

requirements.txt

-# runtime requirements
-numpy>=1.23.5
-tqdm>=4.62.3
-typing_extensions>=4.10.0
+# Runtime requirements
 cloudpickle
-ml_dtypes
+ml-dtypes
+numpy>=1.23.5
 psutil
 torch
+torch>=2.7; platform_system == 'Darwin'
+tqdm>=4.62.3
+typing-extensions>=4.10.0

testing/conftest.py

+import os
+import random
+
+os.environ["PYTHONHASHSEED"] = "0"
+
+random.seed(0)
+
+try:
+    import torch
+except ImportError:
+    pass
+else:
+    torch.manual_seed(0)
+
+try:
+    import numpy as np
+except ImportError:
+    pass
+else:
+    np.random.seed(0)