pe-util VS v6d

FROM fedora:latest # # Set up system # WORKDIR /root RUN dnf -y update RUN dnf clean all RUN dnf install -y git cmake file gcc make man sudo tar RUN dnf install -y gcc-c++ boost boost-devel # # Build peldd to find dlls of exes # RUN git clone https://github.com/gsauthof/pe-util WORKDIR pe-util RUN git submodule update --init RUN mkdir build WORKDIR build RUN cmake .. -DCMAKE_BUILD_TYPE=Release RUN make RUN mv /root/pe-util/build/peldd /usr/bin/peldd RUN chmod +x /usr/bin/peldd # # Add package.sh # ADD package.sh /usr/bin/package.sh RUN chmod +x /usr/bin/package.sh # # Install Windows libraries # RUN dnf install -y mingw64-gcc RUN dnf install -y mingw64-freetype RUN dnf install -y mingw64-cairo RUN dnf install -y mingw64-harfbuzz RUN dnf install -y mingw64-pango RUN dnf install -y mingw64-poppler RUN dnf install -y mingw64-gtk3 RUN dnf install -y mingw64-winpthreads-static RUN dnf install -y mingw64-glib2-static # # Install rust # RUN useradd -ms /bin/bash rustacean USER rustacean RUN curl https://sh.rustup.rs -sSf | sh -s -- -y RUN /home/rustacean/.cargo/bin/rustup update # # Set up rust for cross compiling # RUN /home/rustacean/.cargo/bin/rustup target add x86_64-pc-windows-gnu ADD cargo.config /home/rustacean/.cargo/config ENV PKG_CONFIG_ALLOW_CROSS=1 ENV PKG_CONFIG_PATH=/usr/x86_64-w64-mingw32/sys-root/mingw/lib/pkgconfig/ ENV GTK_INSTALL_PATH=/usr/x86_64-w64-mingw32/sys-root/mingw/ # # Setup the mount point # VOLUME /home/rustacean/src WORKDIR /home/rustacean/src # # Build and package executable # CMD ["/usr/bin/package.sh"]

Brief Overview for any interested: Vineyard (v6d) is an in-memory immutable data manager that provides out-of-the-box high-level abstraction and zero-copy in-memory sharing for distributed data in big data tasks, such as graph analytics (e.g., GraphScope), numerical computing (e.g., Mars), and machine learning.

- https://github.com/v6d-io/v6d/pull/948

We propose vineyard, https://github.com/v6d-io/v6d to address such challenges, which, provides efficient zero-copy data sharing between different compute engines, without extra cost of copying and serialization, compared other similar solutions.