Add AMD OpenCL kernel, runtime-loaded CUDA, mixed backend, portability

AMD GPU backend:
- Add the GCN-tuned equihash192_7.cl kernel (clearCounter/blake/round1..7/
  combine pipeline) and its host driver src/gpu_amd.rs. GpuSolver now dispatches
  AMD-vendor OpenCL devices to it and other devices to the existing kernel
  (force with ZCL_OPENCL_KERNEL=amd|legacy). Validated on an RX 9060 XT: GPU
  solutions match the CPU reference 1/1.
- Expose BatchHasher::midstate() for the kernel's ulong8 hashState arg.

Runtime-loaded GPU drivers (minimum host deps):
- dlopen libcuda / libnvidia-ml via libloading instead of linking them
  (src/dylib.rs macro; cuda.rs, nvml.rs, gpu_probe.rs). The binary now builds
  and starts on hosts without an NVIDIA driver and reports no CUDA devices
  gracefully; remove build.rs (its only job was linking those libs).
- Add Dockerfile.portable + build-portable.sh: build against Debian bullseye's
  glibc 2.31 for a binary that runs on older distros and drives both AMD
  (OpenCL) and NVIDIA (CUDA) cards. Document the build matrix in the README.

Mixed backend (default):
- Add --backend mixed (now the default): each card on its native backend
  (NVIDIA->CUDA, AMD/Intel->OpenCL), deduped so no card is mined twice.
  --devices indexes the unified list shown by --list-devices.

Misc:
- Stale-work timeout (--job-timeout) default 300s -> 600s (10 minutes).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Cargo.toml

@@ -23,13 +23,18 @@ socket2 = "0.5"
 ocl = { version = "0.19", optional = true }
 ratatui = "0.30.0"
 eframe = { version = "0.28", optional = true }
+# Runtime loader for the CUDA driver / NVML (dlopen'd, not link-time, so the
+# binary has no build- or load-time dependency on libcuda / libnvidia-ml).
+libloading = { version = "0.8", optional = true }
 
 [features]
 default = ["gpu", "cuda", "config-gui"]
 gpu = ["dep:ocl"]
 # CUDA backend: drives miniZ's embedded Equihash 192,7 fatbin via the CUDA driver
-# API. build.rs only links libcuda (no nvcc / kernel compilation needed).
-cuda = []
+# API. The driver (libcuda) and NVML are dlopen'd at runtime via libloading, so
+# there is no build-time or load-time dependency on them — the binary builds and
+# starts on hosts without an NVIDIA driver and simply reports no CUDA devices.
+cuda = ["dep:libloading"]
 # Optional native GUI config editor (the `jackpotminer-config` binary). Off by
 # default so the miner never pulls in the GUI toolkit.
 config-gui = ["dep:eframe"]