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>

Dockerfile.portable

@@ -0,0 +1,40 @@
+# Portable jackpotminer build.
+#
+# Links against Debian bullseye's glibc 2.31 (released 2020), so the resulting
+# binary runs on essentially any Linux from the last several years instead of
+# requiring the build host's (much newer) glibc. This is the real fix for the
+# "version `GLIBC_2.39' not found" class of errors — static linking can't solve
+# it for a GPU build, because the GPU driver libraries are glibc-only and load
+# at runtime.
+#
+# The CUDA driver and NVML are dlopen'd at runtime (see src/dylib.rs), so this
+# build needs NO NVIDIA toolkit — only the OpenCL ICD loader (to link libOpenCL).
+# The result is one binary that drives AMD cards (OpenCL) and NVIDIA cards (CUDA,
+# loaded if libcuda.so.1 is present at runtime).
+#
+# Build:   DOCKER_BUILDKIT=1 docker build -f Dockerfile.portable \
+#            --output type=local,dest=dist .
+# or just: ./build-portable.sh
+# Output:  dist/jackpotminer
+
+FROM debian:bullseye-slim AS build
+ENV DEBIAN_FRONTEND=noninteractive
+# gcc/g++ for the linker; ocl-icd-opencl-dev provides libOpenCL.so for linking
+# (the runtime host supplies its own libOpenCL.so.1 via its GPU driver).
+RUN apt-get update && apt-get install -y --no-install-recommends \
+        ca-certificates curl gcc g++ make pkg-config ocl-icd-opencl-dev \
+    && rm -rf /var/lib/apt/lists/*
+# Minimal stable Rust toolchain.
+RUN curl -fsSL https://sh.rustup.rs | sh -s -- -y --profile minimal --default-toolchain stable
+ENV PATH="/root/.cargo/bin:${PATH}"
+
+WORKDIR /src
+COPY . .
+# Miner only (no GUI config tool, to avoid pulling X11/Wayland/GL into the
+# build): AMD OpenCL + dlopen'd CUDA. `--locked` keeps it reproducible.
+RUN cargo build --release --locked --no-default-features --features gpu,cuda \
+    && strip target/release/jackpotminer
+
+# Export just the binary to the build output directory.
+FROM scratch AS export
+COPY --from=build /src/target/release/jackpotminer /jackpotminer