The AMD Open64 SDK is a set of tools, libraries, documentation and headers that developers can use to create high performing applications that run on Linux® operating systems.
- World class code generation tools with advanced optimizations and multi-threading features
- Full native support for OpenMP directive or multi-core parallelization with C, C++, and Fortran languages
- Auto-parallelization for the latest AMD microprocessors
- Graphical performance profiler to monitor and measure application performance
- Optimized math libraries that provide an easy path to faster, multi-threaded applications
- Support for various MPI implementations, e.g. OpenMPI, MPICH1, MPICH2, LAM, and more
- Compatible with GNU toolchain
- Host Processor: 64-bit or 32-bit x86 processor based machine with one or more AMD microprocessors.
- Operating System: The software must be co-installed with a version of the Linux operating system with 32-bit or 64-bit kernel version 2.4.18 or higher. Supported distributions are SLES 11 SP1 and RHEL 5.5 Building 64-bit executables requires a 64-bit operating system.
- Memory requirements: 16 MB or more
- Hard disk: 250 MB during installation
- Other: Web browser and Adobe Acrobat Reader for viewing online documentation.
- bunzip2 and untar the software in some temporary place OR run following command to extract the contents, tar jxvf < filename>
- Change to the top directory
- Run the installation script as root “./install.sh”
- Adjust $PATH settings and other variables as described in ReadMe.txt
The following versions have been tested together to form this bundle. These versions may not reflect the latest releases of each individual component. For more information about the latest releases, please check each component’s product page.
AMD Open64 SDK 1.0
- x86 Open64 Compiler Suite 4.2.4
- AMD CodeAnalyst Performance Analyzer 2.9.18
- AMD Core Math Library (ACML) 4.4.0
- AMD LibM 2.1
- AMD Core Math Library for Graphic Processors (ACML-GPU) 1.1.1
x86 Open64 Compiler Suite
The x86 Open64 compiler system is a high performance, production quality code generation tool designed for high performance parallel computing workloads. This suite offers advanced optimizations, multi-threading features, and support for vectorization, interprocedural analysis, loop transformations, and other functions to help accelerate development and tuning of C, C++, and Fortran applications targeting 32-bit and 64-bit Linux platforms.
Compiler optimizations include:
- Inter-procedural optimization support function inlining, including library functions
- Full support for profile feedback based compiler optimizations
- Vectorization support (SSE4)
- Memory hierarchy and memory allocation optimizations, including support for 1GB huge pages
- Loop unrolling, loop fusion, and cache tiling
- Enhanced auto-parallelization of loops specifically optimized for multi-core processors
- State-of-the-art dependence analysis and global optimization
- Support for processor affinity mapping in the OpenMP and parallel runtime library
- Multi-core scalability optimizations enabled with compiler option
AMD CodeAnalyst Performance Analyzer
AMD CodeAnalyst Performance Analyzer is an open source, front-end graphical user interface to Oprofile. AMD CodeAnalyst taps into the performance monitoring capabilities of AMD processors and supports the collection of up to 32 performance events in a single test run. This capability makes it easier to stage performance experiments. AMD CodeAnalyst also supports Instruction-Based Sampling (IBS), a new performance measurement technique available on AMD Family 10h processors.
AMD Core Math Library (ACML)
ACML provides a free set of thoroughly optimized and threaded math routines for HPC, scientific, engineering and related compute-intensive applications. ACML is ideal for weather modeling, computational fluid dynamics, financial analysis, oil and gas applications, and more. Included are optimized BLAS and LAPACK routines, Fast Fourier Transforms (FFTs), scalar/vector/array math transcendental library routines, and Random Number Generators. ACML is OpenMP enabled.
AMD LibM is a software library containing a collection of basic math functions optimized for x86-64 processor based machines. AMD LibM is a C library, which users can link into their applications to replace compiler-provided math functions. Included in LibM 2.1 are 46 vector functions.