unsloth/tests/studio/install
Matt Van Horn 08c3878919
fix: use partial hipinfo output on crash to avoid CPU fallback (RDNA 4 / gfx1200) (#6292)
* fix: use partial hipinfo output on crash to avoid CPU fallback (#6043)

`hipinfo.exe` on some RDNA 4 hosts (e.g. RX 9060 XT / gfx1200) exits
with STATUS_ACCESS_VIOLATION (0xC0000005) after printing the
gcnArchName line.  The previous guard `$LASTEXITCODE -eq 0` in
studio/setup.ps1 and `if result.returncode == 0` in
install_python_stack.py discarded this partial-but-valid output,
causing the installer to fall through to WMI name inference which sets
HasROCm=false and installs CPU PyTorch instead of the ROCm wheel.

Fix: check for gcnArchName in stdout first; accept the arch regardless
of exit code.  Only fall through to the amd-smi / WMI path when no
gcnArchName is present at all (crash before any output, or a genuine
"no device" error).  A cyan INFO substep is emitted when the arch is
recovered from a crashed hipinfo run so users can see what happened.

Adds a regression test covering the crash-with-valid-output path.

Fixes #6043

* Fix/adjust hipinfo crash fallback for PR #6292

---------

Co-authored-by: Matt Van Horn <455140+mvanhorn@users.noreply.github.com>
Co-authored-by: wasimysaid <wasimysdev@gmail.com>
Co-authored-by: Lee Jackson <130007945+Imagineer99@users.noreply.github.com>
2026-06-15 13:26:04 +02:00
..
conftest.py fix/strix halo and windows AMD ROCm support (#5301) 2026-05-29 22:29:56 -07:00
smoke_test_llama_prebuilt.py Formatting: ruff line-length 100, kwarg-spacing passes, drop blank after short local imports (#6079) 2026-06-08 04:24:13 -07:00
smoke_test_parallel_studio_home.py Reduce and tighten code comments and docstrings repo-wide (#6095) 2026-06-08 23:09:51 -07:00
test_cuda_repair.py Installer: GPU detection follow-ups after #6174 (poisoned venv repair, llama.cpp routing, probe bounds) (#6183) 2026-06-11 05:06:02 -07:00
test_gpu_detection_followups.py Installer: GPU detection follow-ups after #6174 (poisoned venv repair, llama.cpp routing, probe bounds) (#6183) 2026-06-11 05:06:02 -07:00
test_hf_auth.py Studio CI: stop HF 429 rate limits from sinking the llama.cpp prebuilt path (#6199) 2026-06-11 06:57:48 -07:00
test_install_llama_prebuilt_logic.py Installer: drop the lemonade ROCm fallback now the fork ships identical per-gfx prebuilts (#6225) 2026-06-12 11:53:26 -03:00
test_launch_studio_launcher.py Installer: drop redundant -WindowStyle Hidden from the Windows launcher VBS (#6284) 2026-06-12 23:57:30 -07:00
test_llama_pr_force_and_source.py Installer: drop the lemonade ROCm fallback now the fork ships identical per-gfx prebuilts (#6225) 2026-06-12 11:53:26 -03:00
test_macos_version_compat.py Formatting: ruff line-length 100, kwarg-spacing passes, drop blank after short local imports (#6079) 2026-06-08 04:24:13 -07:00
test_pr4562_bugfixes.py Source llama.cpp prebuilts from unslothai/llama.cpp (CUDA, ROCm, macOS) (#5963) 2026-06-10 08:49:57 -07:00
test_pr5940_followups.py Windows/WSL installer: fix winget msstore cert failure, amd-smi DiskPart prompt, and enable AMD GPU (Strix Halo gfx1151) (#5940) 2026-06-10 04:24:49 -07:00
test_probe_timeouts.py Installer: GPU detection follow-ups after #6174 (poisoned venv repair, llama.cpp routing, probe bounds) (#6183) 2026-06-11 05:06:02 -07:00
test_rocm_support.py fix: use partial hipinfo output on crash to avoid CPU fallback (RDNA 4 / gfx1200) (#6292) 2026-06-15 13:26:04 +02:00
test_selection_logic.py Studio: don't silently fall back to a CPU prebuilt on NVIDIA Linux GPU hosts (#6310) 2026-06-14 02:06:31 -03:00