Releases: xkszltl/Roaster
Releases · xkszltl/Roaster
Roaster docker 20220427
New features
- More supported distros:
- CentOS 7.9.2009
- Debian 11.3 (experimental with limited CUDA ecosystem, partially borrowed from Ubuntu 22.04)
- Ubuntu 18.04.6
- Dedicated DEB layers for large packages with higher concurrency of downloading.
- Python multilib installation.
- Build for all version of Python on system when possible.
- Pick different lib versions depending on python/lib coupling.
- Semi-automatic release note generation.
Docker collection
roasterproject/centos:20220427
[-36b7bf1509f1
]roasterproject/debian:20220427
[-7c683c9aaad3
]roasterproject/ubuntu:20220427
[-4d719bb42be7
]
Hash in tag is optional.
Size (gzipped/uncompressed)
- CentOS: 42/103 GiB, 95.3% space efficiency
- Debian: 31/72 GiB, 99.1% space efficiency
- Ubuntu: 30/70 GiB, 98.9% space efficiency
Roaster info
- Version: v8.7.1
Highlights
- GCC
- CentOS: 11.2.1/10.2.1/9.3.1/8.3.1/7.3.1 (SCL devtoolset) and 4.8.5
- Debian: 10.2.1/9.3.0
- Ubuntu: 8.4.0/7.5.0/6.5.0/5.5.0
- Python
- CentOS: 3.8.11/3.6.12 (SCL rh-python) and 3.6.8
- Debian: 3.9.2
- Ubuntu: 3.6.9
- LLVM 14.0.1
- CMake 3.21.6
- CCache 4.6
- CUDA 11.6.2
- cuDNN 8.4.0.27
- TensorRT 8.2.4
- NCCL 2.12.10
- OpenMPI 4.1.3
- UCX 1.12.1
- Intel MPI 2021.5.1
- MKL 2022.0.2
- Eigen 3.4.0
- OpenBLAS 0.3.20
- Boost 1.79.0
- Protobuf 3.19.4
- Pybind 2.9.2
- gRPC 1.45.2
- RocksDB 7.2.0
- TexLive 2022.04.26
- Qt
- CentOS: 5.9.7
- Debian: 5.15.2
- Ubuntu: 5.9.5
- OpenCV 4.5.5
- ONNX 1.11.0+ (main)
- PyTorch 1.11+ (master)
- Fat bin of Pascal/Volta/Tuning included.
- Torchvision/apex included
- Ubuntu 18.04 only has PyTorch 1.10+ because it dropped Python 3.6 support somewhere in-between.
- CentOS version is built with SCL (rh-python38) only and does not work with distro-stock Python 3.6.
- ONNXRuntime 1.11.0
- Fat bin of Pascal/Volta/Tuning/Ampere and Kepler PTX included.
Roaster packages
roaster-axel-2.17.11-0.g1a9b9a5.x86_64
roaster-benchmark-1.6.1-0.g0d98dba.x86_64
roaster-boost-1.79.0-0.g41add36.x86_64
roaster-c-ares-1.18.1-0.g2aa086f.x86_64
roaster-catch2-1.7.0-0.gce37f48.x86_64
roaster-ccache-4.6-0.g19fb17c.x86_64
roaster-cmake-3.21.6-0.g78ae759.x86_64
roaster-cuda-samples-11-6-11.6-0.g7db525e.x86_64
roaster-eigen-3.4.0-0.g3147391.x86_64
roaster-gflags-2.2.2-0.ge171aa2.x86_64
roaster-glog-0.4.0-0.g96a2f23.x86_64
roaster-grpc-1.45.2-0.gb39ffcc.x86_64
roaster-gtest-1.11.0-0.ge2239ee.x86_64
roaster-hiredis-1.0.2-1.g80f5428.x86_64
roaster-ipt-2.0.5-0.ge315d53.x86_64
roaster-jemalloc-5.2.1-0.gea6b3e9.x86_64
roaster-jsoncpp-1.9.5-3.g42e892d.x86_64
roaster-leveldb-1.20.1-0.g4fb1468.x86_64
roaster-libgdiplus-6.1-0.g44b44f9.x86_64
roaster-libpng-1.6.36-0.geddf902.x86_64
roaster-llvm-14.0.1-0.gc620539.x86_64
roaster-lmdb-0.9.29-0.g8ad7be2.x86_64
roaster-lz4-1.9.3-0.gd443718.x86_64
roaster-mkl-dnn-2.6-0.g52b5f10.x86_64
roaster-nsync-1.24.0-0.g712e51f.x86_64
roaster-ompi-4.1.3-0.g0a8d17c.x86_64
roaster-onnx-1.3.0-932.g8ab1efa.x86_64
roaster-onnxruntime-1.11.0-2.gd93c4d6.x86_64
roaster-openblas-0.3.20-0.g0b678b1.x86_64
roaster-opencv-4.5.5-1.gbcc3f10.x86_64
roaster-protobuf-3.19.4-0.g22d0e26.x86_64
roaster-pugixml-1.12.1-0.gdd50fa5.x86_64
roaster-pybind11-2.9.2-0.g914c06f.x86_64
roaster-pytorch-1.8.0.1-12348.gd0d5328.x86_64
roaster-rapidjson-0.0-0.gfcb23c2.x86_64
roaster-rocksdb-7.2.0-0.g14fb86d.x86_64
roaster-shadowsocks-3.3.5-37.gb38d8c9.x86_64
roaster-simdjson-1.0.2-0.g17f3148.x86_64
roaster-snappy-1.1.9-1.g2130738.x86_64
roaster-texlive-2022.04.26-0.g25dd05c.x86_64
roaster-tmux-3.2-0.gbc4681c.x86_64
roaster-ucx-1.12.1-1.ga35c763.x86_64
roaster-utf8proc-2.7.0-0.g8ca6144.x86_64
roaster-vision-0.4.0-2017.g08eab6b.x86_64
roaster-zstd-1.5.2-0.ge47e674.x86_64
Nvidia packages
cuda-cccl-11-6-11.6.55-1.x86_64
cuda-command-line-tools-11-6-11.6.2-1.x86_64
cuda-compat-11-6-510.47.03-1.x86_64
cuda-compiler-11-6-11.6.2-1.x86_64
cuda-cudart-11-6-11.6.55-1.x86_64
cuda-cudart-devel-11-6-11.6.55-1.x86_64
cuda-cuobjdump-11-6-11.6.124-1.x86_64
cuda-cupti-11-6-11.6.124-1.x86_64
cuda-cuxxfilt-11-6-11.6.124-1.x86_64
cuda-documentation-11-6-11.6.124-1.x86_64
cuda-driver-devel-11-6-11.6.55-1.x86_64
cuda-gdb-11-6-11.6.124-1.x86_64
cuda-libraries-11-6-11.6.2-1.x86_64
cuda-libraries-devel-11-6-11.6.2-1.x86_64
cuda-memcheck-11-6-11.6.124-1.x86_64
cuda-nsight-11-6-11.6.124-1.x86_64
cuda-nsight-compute-11-6-11.6.2-1.x86_64
cuda-nsight-systems-11-6-11.6.2-1.x86_64
cuda-nvcc-11-6-11.6.124-1.x86_64
cuda-nvdisasm-11-6-11.6.124-1.x86_64
cuda-nvml-devel-11-6-11.6.55-1.x86_64
cuda-nvprof-11-6-11.6.124-1.x86_64
cuda-nvprune-11-6-11.6.124-1.x86_64
cuda-nvrtc-11-6-11.6.124-1.x86_64
cuda-nvrtc-devel-11-6-11.6.124-1.x86_64
cuda-nvtx-11-6-11.6.124-1.x86_64
cuda-nvvp-11-6-11.6.124-1.x86_64
cuda-samples-11-6-11.6.101-1.x86_64
cuda-sanitizer-11-6-11.6.124-1.x86_64
cuda-toolkit-11-6-11.6.2-1.x86_64
cuda-toolkit-11-6-config-common-11.6.55-1.noarch
cuda-toolkit-11-config-common-11.6.55-1.noarch
cuda-toolkit-config-common-11.6.55-1.noarch
cuda-tools-11-6-11.6.2-1.x86_64
cuda-visual-tools-11-6-11.6.2-1.x86_64
libcudnn8-8.4.0.27-1.cuda11.6.x86_64
libcudnn8-devel-8.4.0.27-1.cuda11.6.x86_64
libnvinfer8-8.2.4-1.cuda11.4.x86_64
libnvinfer-devel-8.2.4-1.cuda11.4.x86_64
libnvinfer-plugin8-8.2.4-1.cuda11.4.x86_64
libnvinfer-plugin-devel-8.2.4-1.cuda11.4.x86_64
libnccl-2.12.10-1+cuda11.6.x86_64
libnccl-devel-2.12.10-1+cuda11.6.x86_64
libnccl-static-2.12.10-1+cuda11.6.x86_64
Pending upstream issues
Only recent and directly build-related issues from Roaster are listed here.
- ccache/ccache#1045
- llvm/llvm-project#54951
- llvm/llvm-project#52820
- microsoft/onnxruntime#11253
- microsoft/onnxruntime#11129
- microsoft/onnxruntime#7597
- mono/libgdiplus#738
- mono/libgdiplus#737
- mono/libgdiplus#723
- pytorch/pytorch#76105
- pytorch/pytorch#76103
- pytorch/pytorch#75496
- pytorch/pytorch#75186
- pytorch/pytorch#73074
- pytorch/pytorch#72393
- pytorch/pytorch#70160
- pytorch/pytorch#69360
- pytorch/pytorch#67284
- pytorch/pytorch#61261
- pytorch/pytorch#58054
- pytorch/vision#5490
- Tencent/rapidjson#2022
TensorRT
Backup for TensorRT 7.2.3.4/8.0.1.6.
Unauthorized use is not allowed.
Please use Nvidia's official distribution and accept their license: https://developer.nvidia.com/tensorrt