Skip to content

Latest commit

 

History

History
209 lines (156 loc) · 8.47 KB

README.md

File metadata and controls

209 lines (156 loc) · 8.47 KB
Raw

SYCLomatic

Introduction

SYCLomatic is a project to assist developers in migrating their existing code written in different programming languages to the SYCL* C++ heterogeneous programming model. Final code editing and verification is a manual process done by the developer.

Use c2s(SYCLomatic binary) command to make it as easy as possible to migrate existing CUDA* codebases to SYCL, which is an industry standard. Once code is migrated to SYCL, it can be compiled and executed by any compiler that implements the SYCL specification as shown here: https://www.khronos.org/sycl/

Releases

The development is in the SYCLomatic branch. Daily builds of the SYCLomatic branch on Linux and Windows* are available at releases. A few times a year, we publish Release Notes to highlight all important changes made in the project: features implemented and issues addressed. The corresponding builds can be found using search in daily releases. None of the branches in the project are stable or rigorously tested for production quality control, so the quality of these releases is expected to be similar to the daily releases.

SYCLomatic supports migrating programs implemented with CUDA versions 8.0, 9.x, 10.x, 11.x, 12.0-12.6. The list of supported languages and versions may be extended in the future.

Build from source code

Prerequisites

  • git - Download
  • cmake version 3.14 or later - Download
  • python 3 - Download
  • ninja - Download
  • C++ compiler
    • Linux: GCC version 7.5.0 or later (including libstdc++) - Download
    • Windows: Visual Studio 2019 or 2022 (In the following description, assume that the version used is 2019) - Download

Note: SYCLomatic can be built from source without any CUDA dependencies. However, before migration of CUDA codebases to SYCL, ensure that CUDA header files are accessible to the tool. These header files are necessary for SYCLomatic to properly understand and process CUDA code during the migration process.

Create SYCLomatic workspace

Throughout this document SYCLOMATIC_HOME denotes the path to the local directory created as SYCLomatic workspace. It might be useful to create an environment variable with the same name.

Linux:

export SYCLOMATIC_HOME=~/workspace
export PATH_TO_C2S_INSTALL_FOLDER=~/workspace/c2s_install
mkdir $SYCLOMATIC_HOME
cd $SYCLOMATIC_HOME

git clone https://github.com/oneapi-src/SYCLomatic.git

Windows (64-bit):

Open a developer command prompt using one of two methods:

  • Click start menu and search for "x64 Native Tools Command Prompt for VS 2019".
  • Win-R, type "cmd", click enter, then run "C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Auxiliary\Build\vcvarsall.bat" x64
set SYCLOMATIC_HOME=%USERPROFILE%\workspace
set PATH_TO_C2S_INSTALL_FOLDER=%USERPROFILE%\workspace\c2s_install
mkdir %SYCLOMATIC_HOME%
cd %SYCLOMATIC_HOME%

git clone https://github.com/oneapi-src/SYCLomatic.git

Build SYCLomatic

Linux:

cd $SYCLOMATIC_HOME
mkdir build
cd build
cmake -G Ninja -DCMAKE_INSTALL_PREFIX=$PATH_TO_C2S_INSTALL_FOLDER  -DCMAKE_BUILD_TYPE=Release  -DLLVM_ENABLE_PROJECTS="clang"  -DLLVM_TARGETS_TO_BUILD="X86;NVPTX" ../SYCLomatic/llvm
ninja install-c2s

Windows (64-bit):

cd %SYCLOMATIC_HOME%
mkdir build
cd build
cmake -G Ninja -DCMAKE_INSTALL_PREFIX=%PATH_TO_C2S_INSTALL_FOLDER%  -DCMAKE_BUILD_TYPE=Release  -DLLVM_ENABLE_PROJECTS="clang"  -DLLVM_TARGETS_TO_BUILD="X86;NVPTX" ..\SYCLomatic\llvm
ninja install-c2s

Note: For build system with 16G or less memory, to avoid out of memory issues, it is recommended to add the option -DLLVM_PARALLEL_LINK_JOBS=1 in the CMake step and use 4 or less total parallel jobs when building SYCLomatic by invoking ninja with the option -j4:

ninja -j4 install-c2s

Build success message:

After a successful build, you should be able to see following message in the terminal output.

-- Install configuration: "Release"
-- Installing: ... bin/dpct
-- Creating c2s

Deployment

Linux:

source $PATH_TO_C2S_INSTALL_FOLDER/setvars.sh

Windows (64-bit):

.\%PATH_TO_C2S_INSTALL_FOLDER%\setvars.bat

Verify the Installation

Before running SYCLomatic, it is important to verify that dpct/c2s(SYCLomatic binary) is installed correctly and PATH is set properly.

To do this, try running the following command and checking if dpct is available after the installation.

c2s --version

Note: If c2s cannot be found, please check the setup of PATH.

Run SYCLomatic

Run c2s command

Get c2s help information by running "c2s --help". dpct is an alias command for c2s.

Test SYCLomatic

Run in-tree LIT tests

SYCLomatic uses LLVM Integrated Tester infrastructure to do the unit test. Note: Certain CUDA header files may need to be accessible to the tool. After building the SYCLomatic, you can run the lit test by:

ninja check-clang-c2s

Or enabling extra lit build test besides lit migration test:

Set up compiler and libraries according to oneAPI Development Environment Setup

export BUILD_LIT=TRUE
ninja check-clang-c2s

Run end to end test suite (Recommend for contributors)

The end to end test executes the migration, build and run steps. Follow instructions from the link below to build and run tests: README

Known Issues and Limitations

  • SYCL* 2020 support work is in progress.
  • 32-bit host/target is not supported.

Useful Links

License

See LICENSE for details.

Contributing

See CONTRIBUTING.md for details.

Trademarks information

Intel, the Intel logo, and other Intel marks are trademarks of Intel Corporation or its subsidiaries.
*Other names and brands may be claimed as the property of others. SYCL is a trademark of the Khronos Group Inc.