PKGBUILD's for building NVIDIA's RAPIDS (cuML/cuDF) on Arch Linux + CUDA11 a procedure of building NVIDIA's RAPIDS under Arch Linux with CUDA environment on '''Single Nvidia's GPU'''.
- Hardware
** Intel or AMD's CPU
** NVIDIA's Single GPU which has architecture whose type is Pascal / Volta / Turing / Ampere. (eg GeForce / TITAN / Tesla / Quadro)
- Software
** Arch Linux or its derivatives.
** AUR / yay required
** Build on CUDA11.1 (+CUDNN8 + NCCL)
- Disclaimer
Operation is not necessarily guaranteed. The author are not responsible for any damage of your environment by any of the operations described here.
* Set ver 3.12.4 (do not build ver 3.13 or higher because its incompatible with cuDF).
* static library required for building treelite
* Set ver 0.17.1 (do not build ver 1.0.0 or higher because its incompatible with cuDF).
* Build with "CUDA" (differencet from AUR).
* Apache Orc is required located in AUR.
* During compilation in Apache Flight, there is a patch so that gRPC-c++-plugin cab be easily integrated in Higher Version.
* Build Python Library for enabling to run cmake during "python setup.py build"
* An implementation of NumPy-compatible array on CUDA presented by Preferred Networks.
* AUR provides the same package. However, AUR only contains older ver (7.2.0). which does not support CUDA 11.0 , (current ver. is 8.0.0)
* See [github](https://github.com/cupy/cupy)
* an optimized communication layer for Message Passing (MPI), PGAS/OpenSHMEM libraries and RPC/data-centric applications. and wrapper for python.
* See [OpenUCX's https://github.com/openucx/ucx)
* See [ucx-py's github](https://github.com/rapidsai/ucx-py)
* an optimized distributed gradient boosting library
* Builds Library and its Python-Wrapper
* See [github](https://github.com/dmlc/xgboost)
* a model compiler for efficient deployment of decision tree ensembles
* Builds Library and its Python-Wrapper
* See [github](https://github.com/dmlc/treelite)
8 RMM
* RAPIDS Memory Manager provided by Rapids.
* See [github](https://github.com/rapidsai/rmm)
* amend PKGBUILD and setup.py.patch to avoid missing libraries (cudart fmt...) (2020.8.21. 2300GMT)
* Utilities for CUDA and DASK.
* See [github](https://github.com/rapidsai/dask-cuda)
10 nvtx, python-nvtx
* NVIDIA Tool Extension Library
11 cuDF
* a GPU DataFrame taking place of pandas
* See [github](https://github.com/rapidsai/cudf)
12 cuML
* A suite of libraries that implement machine learning algorithms presented by Rapids.
* See [github](https://github.com/rapidsai/cuml)
* amend PKGBUILD and setup.py.patch to avoid missing libraries (cublas cusuparse cusolver ...) (2020.9.19. 1400GMT)
* required for building cuGraph
* GPU accelerated signal processing which may replace scipy signal?
* See [github](https://github.com/rapidsai/cusignal)
* a bug-fix patch required to build gdal under poppler upgraded to 20.08.1
* a library for handling spatial and trajectory data.
* gdal and python-gdal are needed
* See [github](https://github.com/rapidsai/cuspatial)
* a library for Graph Analytics
* See [github](https://github.com/rapidsai/cugraph)
* cuxfilter ( ku-cross-filter ) is a RAPIDS framework to connect web visualizations to GPU accelerated crossfiltering.
* See [github](https://github.com/rapidsai/cuxfilter)