Fix AMD GPU false detection, blackbox deadlock, and NOGPU build bloat

- sat.go: DetectGPUVendor lspci fallback now checks GPU device classes
  ([0300]/[0302]/[0380]) per line instead of scanning the whole output for
  vendor name; AMD EPYC servers have dozens of AMD-branded PCIe entries
  (Root Complex, IOMMU, Host Bridge) that were triggering the old check
- blackbox.go: fix deadlock in finishCycle — it held w.mu while calling
  persistState(), which acquires rt.mu then re-acquires w.mu inside
  persistStateLocked(); now w.mu is released before persistState()
- build.sh: remove NVIDIA-specific overlay files (bee-gpu-burn,
  bee-john-gpu-stress, bee-nccl-gpu-stress, bee-nvidia-recover,
  bee-dcgmproftester-staggered, bee-check-nvswitch,
  nvidia-fabricmanager.service.d/) for non-nvidia build variants
- bee-selfheal: gate NVIDIA recovery on BEE_GPU_VENDOR=nvidia so the
  script does not attempt to restart bee-nvidia.service on NOGPU builds

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

iso/builder/build.sh

@@ -1419,6 +1419,13 @@ rm -rf \
 if [ "$BEE_GPU_VENDOR" != "nvidia" ]; then
     rm -f "${OVERLAY_STAGE_DIR}/usr/local/bin/bee-nvidia-load"
     rm -f "${OVERLAY_STAGE_DIR}/etc/systemd/system/bee-nvidia.service"
+    rm -f "${OVERLAY_STAGE_DIR}/usr/local/bin/bee-gpu-burn"
+    rm -f "${OVERLAY_STAGE_DIR}/usr/local/bin/bee-john-gpu-stress"
+    rm -f "${OVERLAY_STAGE_DIR}/usr/local/bin/bee-nccl-gpu-stress"
+    rm -f "${OVERLAY_STAGE_DIR}/usr/local/bin/bee-nvidia-recover"
+    rm -f "${OVERLAY_STAGE_DIR}/usr/local/bin/bee-dcgmproftester-staggered"
+    rm -f "${OVERLAY_STAGE_DIR}/usr/local/bin/bee-check-nvswitch"
+    rm -rf "${OVERLAY_STAGE_DIR}/etc/systemd/system/nvidia-fabricmanager.service.d"
 fi
 
 # --- inject authorized_keys for SSH access ---