=== Fedora ↔ PyTorch Local Check Summary ===
Repo: /home/anandy/pytorch
Conda env: torch-fedora
Report dir: /home/anandy/pytorch-fedora-report

---- Toolchain ----
Python: Python 3.12.11
GCC: gcc (GCC) 15.1.1 20250521 (Red Hat 15.1.1-2)
Clang: clang version 20.1.8 (Fedora 20.1.8-1.fc42)

---- Build warnings (top lines) ----
GCC:
      7  Unused direct dependencies:
      6 1641:11: warning: ‘void* memcpy(void*, const void*, size_t)’ reading between 8 and 17179869176 bytes from a region of size 0 [-Wstringop-overread]
      6 1641:11: warning: ‘void* memcpy(void*, const void*, size_t)’ reading between 4 and 8589934588 bytes from a region of size 0 [-Wstringop-overread]
      4 192:9: warning: ‘throw’ will always call ‘terminate’ [-Wterminate]
      4 185:9: warning: ‘throw’ will always call ‘terminate’ [-Wterminate]
      4 174:5: warning: ‘throw’ will always call ‘terminate’ [-Wterminate]
      4 160:5: warning: ‘throw’ will always call ‘terminate’ [-Wterminate]
      2 1641:11: warning: ‘void* memcpy(void*, const void*, size_t)’ reading between 1 and 2147483647 bytes from a region of size 0 [-Wstringop-overread]
      1 90: DevelopDeprecationWarning: develop command is deprecated.
      1 82: SetuptoolsDeprecationWarning: `project.license` as a TOML table is deprecated
      1 41: SyntaxWarning: invalid escape sequence '\('

Clang:
      1 90: DevelopDeprecationWarning: develop command is deprecated.
      1 82: SetuptoolsDeprecationWarning: `project.license` as a TOML table is deprecated

---- Smoke test ----
Torch: 2.9.0a0+git47a1db8
Platform: Linux-6.15.4-200.fc42.x86_64-x86_64-with-glibc2.41
OpenMP available: True
Matmul OK: True

---- /dev/shm & DataLoader ----
Filesystem      Size  Used Avail Use% Mounted on
tmpfs            16G  238M   16G   2% /dev/shm

/dev/shm total ~15759 MB
Batches processed: 50
Batches processed: 100
DataLoader stress: done, total batches: 147

---- Distributed TCPStore ----
TCPStore failed: module 'torch.distributed' has no attribute 'init_process_group'

---- CUDA probe ----
nvidia-smi: None
torch.version.cuda: None
cuda.is_available(): False

==== Suggested PRs (based on findings) ====
- Distributed: Improve TCPStore diagnostics & firewall hint (firewalld) and small retry on bind.
- Warnings: Submit focused cleanups for top GCC/Clang warnings (correctness-preserving casts, sign-compare).