/llvm-scripts

Script to install different versions of Clang/LLVM

Primary LanguageShell

License

Please see LICENSE for usage terms.

Introduction

Archer Logo

Archer is a data race detector for OpenMP programs.

Archer combines static and dynamic techniques to identify data races in large OpenMP applications, leading to low runtime and memory overheads, while still offering high accuracy and precision. It builds on open-source tools infrastructure such as LLVM, ThreadSanitizer, and OMPT to provide portability.

Prerequisites

To compile Archer you need an host Clang/LLVM version >= 3.9, a CMake version >= 3.4.3.

Ninja build system is preferred. For more information how to obtain Ninja visit https://martine.github.io/ninja.

Archer has been tested with the LLVM OpenMP Runtime version >= 3.9, and with the LLVM OpenMP Runtime with OMPT support currently under development at https://github.com/OpenMPToolsInterface/LLVM-openmp (under the branch "align-to-tr").

Installation

Archer has been developed under LLVM 3.9 (for more information visit http://llvm.org).

Manual Building

For a manual building please visit the GitHub page https://github.com/PRUNERS/archer.

Automatic Building

Archer comes both as standalone and LLVM tool.

In order to obtain and automatically build Clang/LLVM with Archer support execute the following commands in your command-line (instructions are based on bash shell, GCC-4.9.3 version and Ninja build system).

Build Clang/LLVM 3.9 with Archer support by running install.sh:

export LLVM_INSTALL=$HOME/usr
./install.sh <path-to-installation-folder>

The installation script will create a folder called LLVM at the same level of the llvm_archer directory and install LLVM into LLVM_INSTALL. By default the script will try to install the software under "/usr".

Once the installation completes, you need to setup your environement to allow Archer to work correctly.

Please set the following path variables:

export PATH=${LLVM_INSTALL}/bin:${PATH}"
export LD_LIBRARY_PATH=${LLVM_INSTALL}/lib:${LD_LIBRARY_PATH}"

To make the environment permanent add the previous lines or equivalents to your shell start-up script such as "~/.bashrc".

Options

Running the command:

./install --help

shows the options available for building and installing Clang/LLVM with Archer support.

Usage

  ./install.sh [options]

Options
  --prefix=<value>             = Specify an installation path.
  --build-system=<value>       = Specify a build system generator. Please run
                                 'man cmake-generators' for a list of generators
                                 available for this platform. Default is Ninja.
  --release=<value>            = Specify the release version of Clang/LLVM that
                                 will be installed (>= 39). Default is 3.9.
  --http                       = Enables GitHub web url in case SSH key and
                                 passphrase are not set in the GitHub account.
  --llvm-only                  = Build and install only Clang/LLVM without
                                 Archer support and OpenMP runtime.
  --update                     = Update previous building. Default is SSH.
  --omp-tsan-support           = Enabled ThreadSanitizer support in official
                                 LLVM OpenMP runtime. Default is an LLVM OpenMP
                                 Runtime with OMPT support.
  --build-type=<value>         = Specify the type of build. Accepted values
                                 are Release (default), Debug or RelWithDebInfo.
  --gcc-toolchain-path=<value> = Specify the GCC toolchain path.

Publications

  • S. Atzeni, G. Gopalakrishnan, Z. Rakamaric, D. H. Ahn, I. Laguna, M. Schulz, G. L. Lee, J. Protze, and M. S. Müller. 2016. "ARCHER: Effectively Spotting Data Races in Large Openmp Applications." In 2016 IEEE International Parallel and Distributed Processing Symposium (IPDPS), 53–62. http://ieeexplore.ieee.org/document/7516001/
  • J. Protze, S. Atzeni, D. H. Ahn, M. Schulz, G. Gopalakrishnan, M. S. Müller, I. Laguna, Z. Rakamarić, and G. L. Lee. 2014. "Towards Providing Low-Overhead Data Race Detection for Large Openmp Applications." In Proceedings of the 2014 LLVM Compiler Infrastructure in HPC, 40–47. http://dl.acm.org/citation.cfm?id=2688369

Contacts and Support

Sponsors

UofU Logo LLNL Logo RWTH AACHEN Logo