# build stage
FROM registry.fedoraproject.org/fedora:rawhide AS builder

# rocm 6.4.4 repo
RUN <<'EOF'
tee /etc/yum.repos.d/rocm.repo <<REPO
[ROCm-6.4.4]
name=ROCm6.4.4
baseurl=https://repo.radeon.com/rocm/el9/6.4.4/main
enabled=1
priority=50
gpgcheck=1
gpgkey=https://repo.radeon.com/rocm/rocm.gpg.key
REPO
EOF

# deps
RUN dnf -y --nodocs --setopt=install_weak_deps=False \
    --exclude='*sdk*' --exclude='*samples*' --exclude='*-doc*' --exclude='*-docs*' \
    install \
      make gcc cmake lld clang clang-devel compiler-rt libcurl-devel ninja-build \
      rocm-llvm rocm-device-libs hip-runtime-amd hip-devel \
      rocblas rocblas-devel hipblas hipblas-devel rocm-cmake libomp-devel libomp \
      rocminfo radeontop \
      git-core vim sudo rsync \
 && dnf clean all && rm -rf /var/cache/dnf/*

# rocm env
ENV ROCM_PATH=/opt/rocm \
    HIP_PATH=/opt/rocm \
    HIP_CLANG_PATH=/opt/rocm/llvm/bin \
    HIP_DEVICE_LIB_PATH=/opt/rocm/amdgcn/bitcode \
    PATH=/opt/rocm/bin:/opt/rocm/llvm/bin:$PATH

# rocWMMA 
WORKDIR /opt
COPY ./build-rocwmma.sh .
RUN chmod +x build-rocwmma.sh && ./build-rocwmma.sh

# llama.cpp
WORKDIR /opt/llama.cpp
RUN git clone --recursive https://github.com/ggerganov/llama.cpp.git .

# Apply # rocWMMA patch
COPY ./apply-rocwmma-fix.sh /opt/apply-rocwmma-fix.sh
RUN chmod +x /opt/apply-rocwmma-fix.sh && /opt/apply-rocwmma-fix.sh /opt/llama.cpp

# build
RUN git clean -xdf \
 && git pull \
 && git submodule update --recursive \
 && echo -e '#ifndef HIP_HAS_SHFL_SYNC_FUNCS\n#define HIP_HAS_SHFL_SYNC_FUNCS\n#ifndef __shfl_sync\n#define __shfl_sync(mask,var,srcLane,width) __shfl(var,srcLane,width)\n#endif\n#ifndef __shfl_xor_sync\n#define __shfl_xor_sync(mask,var,laneMask,width) __shfl_xor(var,laneMask,width)\n#endif\n#ifndef __shfl_up_sync\n#define __shfl_up_sync(mask,var,delta,width) __shfl_up(var,delta,width)\n#endif\n#endif\n' \
    | cat - /opt/llama.cpp/ggml/src/ggml-cuda/mma.cuh > /opt/llama.cpp/ggml/src/ggml-cuda/mma.cuh.tmp \
 && mv /opt/llama.cpp/ggml/src/ggml-cuda/mma.cuh.tmp /opt/llama.cpp/ggml/src/ggml-cuda/mma.cuh \
 && cmake -S . -B build \
      -DGGML_HIP=ON \
      -DAMDGPU_TARGETS=gfx1151 \
      -DCMAKE_BUILD_TYPE=Release \
      -DLLAMA_HIP_UMA=ON \
      -DGGML_HIP_ROCWMMA_FATTN=ON \
      -DROCM_PATH=/opt/rocm \
      -DHIP_PATH=/opt/rocm \
      -DHIP_PLATFORM=amd \
      -DCMAKE_HIP_FLAGS="--rocm-path=/opt/rocm" \
 && cmake --build build --config Release -- -j$(nproc) \
 && cmake --install build --config Release

# libs
RUN find /opt/llama.cpp/build -type f -name 'lib*.so*' -exec cp {} /usr/lib64/ \; \
 && ldconfig

# helper
COPY gguf-vram-estimator.py /usr/local/bin/gguf-vram-estimator.py
RUN chmod +x /usr/local/bin/gguf-vram-estimator.py


# runtime stage
FROM registry.fedoraproject.org/fedora-minimal:rawhide

# rocm 6.4.3 repo
RUN <<'EOF'
tee /etc/yum.repos.d/rocm.repo <<REPO
[ROCm-6.4.4]
name=ROCm6.4.4
baseurl=https://repo.radeon.com/rocm/el9/6.4.4/main
enabled=1
priority=50
gpgcheck=1
gpgkey=https://repo.radeon.com/rocm/rocm.gpg.key
REPO
EOF

# runtime deps
RUN microdnf -y --nodocs --setopt=install_weak_deps=0 \
    --exclude='*sdk*' --exclude='*samples*' --exclude='*-doc*' --exclude='*-docs*' \
    install \
      bash ca-certificates libatomic libstdc++ libgcc libgomp sudo \
      hip-runtime-amd rocblas hipblas \
      rocminfo radeontop \
 && microdnf clean all && rm -rf /var/cache/dnf/*

# copy
COPY --from=builder /usr/local/ /usr/local/

# ld
RUN echo "/usr/local/lib"  > /etc/ld.so.conf.d/local.conf \
 && echo "/usr/local/lib64" >> /etc/ld.so.conf.d/local.conf \
 && ldconfig \
 && cp -n /usr/local/lib/libllama*.so* /usr/lib64/ 2>/dev/null || true \
 && ldconfig

# helper
COPY gguf-vram-estimator.py /usr/local/bin/gguf-vram-estimator.py
RUN chmod +x /usr/local/bin/gguf-vram-estimator.py

# profile
RUN printf '%s\n' \
  'export ROCBLAS_USE_HIPBLASLT=1' \
  > /etc/profile.d/rocm.sh && chmod +x /etc/profile.d/rocm.sh \
  && echo 'source /etc/profile.d/rocm.sh' >> /etc/bashrc

# shell
CMD ["/bin/bash"]