Intel Fortran Compiler

Overview

The compilers generate code for IA-32 and Intel 64 processors and certain non-Intel but compatible processors, such as certain AMD processors. A specific release of the compiler (11.1) remains available for development of Linux-based applications for IA-64 (Itanium 2) processors. On Windows, it is known as Intel Visual Fortran. On OS X and Linux, it is known as Intel Fortran.

The latest release of the compiler continues to support the Intel Xeon Phi coprocessor and Intel Architecture instruction-set capabilities by means of automatic vectorization, which can enable applications to use SSE, SSE2, SSE3, SSSE3, SSE4 and AVX SIMD instructions. Use of such instructions through the compiler can lead to improved application performance in some applications as run on IA-32 and Intel 64 architectures, compared to applications built with compilers that do not support these instructions.

Intel Fortran also continues support for OpenMP 4.0, automatic parallelization for symmetric multiprocessing, almost all of the Fortran 2003 standard and much of the Fortran 2008 standard including Coarray Fortran, user-defined I/O, BLOCK and submodules. For more information on Fortran standards, a number or resources are available, such as the Wikipedia Fortran entry or the Fortran 2008 wiki page. When used with Intel cluster tools (see the "Description of Packaging" below) the compiler can also automatically generate Message Passing Interface calls for distributed memory multiprocessing from OpenMP directives.

Optimizations

Intel compilers are optimized to computer systems using processors that support Intel architectures. They are designed to minimize stalls and to produce code that executes in the fewest possible number of cycles. Intel Fortran Compilers support three separate high-level techniques for optimizing the compiled program: interprocedural optimization (IPO), profile-guided optimization (PGO), and other high-level optimizations (HLO). They also support a directives-based approach to application offloading to Intel coprocessors, such as the Intel Xeon Phi coprocessor.

Interprocedural optimization applies typical compiler optimizations (such as constant propagation) but uses a broader scope that may include multiple procedures, multiple files, or the entire program.

Regarding profile-guided optimization,the compiler generates a dataset of performance-related information from using the application with representative workloads, which it then analyzes to find which parts of the application are executed more and less frequently. The compiler uses these data to organize application execution to optimize performance based on how the application is actually used. This is in contrast to IPO which optimizes applications according to the logical flow of the application independent of workloads. The two can be combined to provide workload-based optimizations within which the logical-flow is optimized. Thus, all optimizations can benefit from profile-guided feedback because they are less reliant on heuristics when making compilation decisions.

High-level optimizations are optimizations performed on a version of the program that more closely represents the source code. This includes loop interchange, loop fusion, loop unrolling, loop distribution, data prefetch, and more.

Standards support

The Intel Fortran compiler supports all of the features of the Fortran 90, Fortran 95, Fortran 2003 standards and most of Fortran 2008. It also supports some draft Fortran 2015 features. Additionally, it supports various extensions found in VAX Fortran and Compaq Visual Fortran.

A partial list of items from the Fortran 2003 standard supported by Intel Fortran:

A partial list of items from the Fortran 2008 standard supported by Intel Fortran:

Fortran 2015, currently a draft version, includes further interoperability between Fortran and C. Intel Fortran supports draft Fortran 2015 as of version 16.0.

Architectures

Description of packaging

The Intel Fortran compiler is available as part of the Intel Parallel Studio XE 2016 suite, which focuses on development of parallelism models in application software. It also includes Intel C++, Intel Math Kernel Library, Intel Integrated Performance Primitives, Intel Data Analytics Acceleration Library and performance analysis tools such as Intel VTune Amplifier and Intel Inspector. There are three forms of Parallel Studio XE: Composer, Professional, and Cluster. The Composer Edition includes the C++ and/or Fortran compilers, the performance libraries, and parallel models support. The Professional Edition adds the analysis tools that assist in debugging and tuning parallel applications. The Cluster Edition adds support for development of software for computer clusters. It includes all of the above plus a standards-based MPI Library, MPI communications profiling and analysis tool, MPI error checking and tuning tools, and cluster checker.

Debugging

The Intel compiler provides debugging information that is standard for the common debuggers (DWARF 2 on Linux, similar to gdb, and COFF for Windows). The flags to compile with debugging information are /Zi on Windows and -g on Linux. Debugging is done on Windows using the Visual Studio debugger, and on Linux using gdb.

While the Intel compiler can generate a gprof-compatible profiling output, Intel also provides a kernel-level, system-wide statistical profiler as a separate product called VTune. VTune features an easy-to-use GUI (integrated into Visual Studio for Windows, Eclipse for Linux) as well as a command-line interface. In addition to the VTune profiler, there is Intel Advisor that specializes in vectorization optimization and tools for threading design and prototyping.

Intel also offers a tool for memory and threading error detection called Intel Inspector XE. Regarding memory errors, it helps detect memory leaks, memory corruption, allocation/de-allocation of API mismatches and inconsistent memory API usage. Regarding threading errors, it helps detect data races (both heap and stack), deadlocks and thread and synch API errors.