Advanced optimizing compiler for IBM Power Systems running Linux distributions
IBM® XL Fortran for Linux is a standards-based, high-performance Fortran compiler with advanced optimization and performance-tuning features. It enables the development of parallel applications that exploit the latest IBM POWER8® processor with multi-core and vector features. XL Fortran for Linux optimizes your infrastructure on IBM Power Systems™ in support of extensive numerical, scien-tific and high-performance computing.
XL Fortran for Linux provides a single XL Fortran offering with compiler versions for both big endian and little endian architectures. (Little endian only) Supports Linux distributions such as RHEL 7.1, RHEL 7.2, Ubuntu Server 14.04, Ubuntu Server 14.10, Ubuntu Server 16.04, CentOS 7, SLES 12 and SLES 12 SP1. It also supports creating or porting applications for execution on IBM systems built on the POWER8 technology.
IBM XL Fortran for Linux delivers
- Leading-edge optimization and performance tuning technology that takes advantage of hardware advancements in Power Systems™,including the latest IBM POWER8 technology.
- High-performance libraries that are specifically tuned for the POWER® architecture, including IBM Mathematical Acceleration Subsystem and Basic Linear Algebra Subprograms libraries.
- Profile-driven optimization with monitoring code used to generate a runtime profile, helping to tune application performance.
- Language standard compliance for adherence with the latest international programming language standards, including Fortran 2003, a subset of Fortran 2008 and language interoperability standards.
- Support for a commonly used subset of CUDA Fortran to enable applications to exploit the NVIDIA GPU.
Leading-edge optimization and performance tuning technology
- Supports systems with processor architectures that execute single instruction, multiple data (SIMD) instruction sets, including vector multimedia extension (VMX) instructions and the POWER8 vector scalar extension (VSX) instruction set.
- Delivers full support of OpenMP Version 3.1 and partial support for OpenMP Version 4.5 to enable parallel application development and simplify data and task parallelism.
- Supports profile-driven optimization.
- Supports automatic parallelization of sequential programs to optimize the power of multi-core processors.
- The math libraries are highly tuned for the POWER architectures and offer both the scalar and vector versions.
- Mathematical Acceleration Subsystem (MASS) consists of libraries of mathematical intrinsic functions tuned specifically for optimum performance on POWER architectures.
- Basic Linear Algebra Subprograms (BLAS) consists of libraries of high-performance algebraic functions, performing combined matrix multiplication and addition on general matrices or their transposes, and compute the matrix-vector product for a general matrix or its transpose.
- Profile-directed feedback optimization collects information about a program run with typical input data and then applies transformations to the program based on that information.
- Profile-directed feedback can ensure the performance of the application is optimized for its important inputs.
- Application profile monitoring and profile directed feedback capabilities minimize the need for manual tuning to achieve desirable performance on large, complex applications.
Language standard compliance
- Supports FORTRAN 77, Fortran 90, Fortran 95, Fortran 2003 and a subset of Fortran 2008.
- Supports a subset of Technical Specification 29113 for further interoperability of Fortran with C.
- Helps to protect investments in existing Fortran source and object code built with previous versions of the compilers, allowing users to rebuild and link all the components into the same applications using the latest compiler versions.
- Supports compatibility of object code or libraries compiled with previous versions of XL Fortran with the newest XL Fortran compiler and runtime environment.
- Supports industry-compliant Fortran language standards for application portability.
Support for a commonly used subset of CUDA Fortran
- XL Fortran for Linux on little endian distributions, V15.1.4, supports a commonly used subset of the CUDA Fortran constructs.
- Increases the computing performance of many applications by using CUDA directly or by linking to GPU-accelerated libraries.
XL Fortran for Linux resources