12 Detailed Contents

• 1 Introduction
  • 1.1 Introduction to the Compiler
    • 1.1.1 Other Fortran-related Activities at NAG
    • 1.1.2 This Manual
• 2 Using the Compiler
  • 2.1 Usage
  • 2.2 Description
  • 2.3 File Types
  • 2.4 Compiler Options
  • 2.5 Files
  • 2.6 Compilation Messages
  • 2.7 Compiler Limits
  • 2.8 Input/Output Information
  • 2.9 OpenMP Support
  • 2.10 Automatic File Preconnection
  • 2.11 IEEE 754 Arithmetic Support
  • 2.12 Half precision floating-point
  • 2.13 Random Number Algorithm
  • 2.14 Automatic Garbage Collection
  • 2.15 Memory Tracing
  • 2.16 Undefined Variable Detection
  • 2.17 Data Types
  • 2.18 Modules
  • 2.19 Runtime Environment Variables
  • 2.20 Debugging
  • 2.21 Producing a Call Graph
  • 2.22 Dependency Analysis
  • 2.23 Generating Interfaces
  • 2.24 Source File Polishing
  • 2.25 Enhanced Source File Polishing
  • 2.26 Unifying Precision
• 3 Debugging with dbx90
  • 3.1 dbx90 command line
  • 3.2 Description of dbx90
  • 3.3 dbx90 options
  • 3.4 dbx90 Commands
  • 3.5 dbx90 Expressions
    • 3.5.1 Scalar expressions
    • 3.5.2 Array sections
    • 3.5.3 Derived type component specification
  • 3.6 dbx90 Command aliases
  • 3.7 dbx90 limitations
  • 3.8 Example of dbx90
  • 3.9 Troubleshooting dbx90
• 4 Preprocessing with fpp
  • 4.1 Overview of fpp
  • 4.2 fpp command line
  • 4.3 Description of fpp
  • 4.4 fpp options
  • 4.5 Using fpp
    • 4.5.1 Source files
    • 4.5.2 Output
    • 4.5.3 Directives
    • 4.5.4 Macro definition
    • 4.5.5 Including external files
    • 4.5.6 Line number control
    • 4.5.7 Conditional selection of source text
  • 4.6 Preprocessing details
    • 4.6.1 Scope of macro or variable definitions
    • 4.6.2 End of macro definition
    • 4.6.3 Function-like macro definition
    • 4.6.4 Cancelling macro definitions
    • 4.6.5 Conditional source code selection
    • 4.6.6 Including external files
    • 4.6.7 Comments
    • 4.6.8 Macro functions
    • 4.6.9 Macro expression
  • 4.7 fpp diagnostics
• 5 Extensions
  • 5.1 Non-standard Extensions
    • 5.1.1 BOZ literal constants outside DATA statements
    • 5.1.2 Longer Names
    • 5.1.3 Dollar Sign in Names
    • 5.1.4 Input/output endian/format conversion
    • 5.1.5 Elemental BIND(C) procedures
    • 5.1.6 Maximum array rank is 31
  • 5.2 Obsolete Extensions
    • 5.2.1 Byte Sizes
    • 5.2.2 TAB Format
    • 5.2.3 Hollerith Constants
    • 5.2.4 D (debug) lines in Fixed Source Form
    • 5.2.5 Increased Line Length in Fixed Source Form
    • 5.2.6 Increased Maximum Number of Continuation Lines
    • 5.2.7 Intrinsic functions with mixed-kind arguments
    • 5.2.8 ACCESS='APPEND' specifier on OPEN statement
    • 5.2.9 VAX FORTRAN TYPE statement
    • 5.2.10 Auto-skipping NAMELIST input
    • 5.2.11 Legacy Application Support
    • 5.2.12 Mismatched Argument Lists
    • 5.2.13 Double Precision Complex Extensions
• 6 Intrinsic Modules
  • 6.1 Intrinsic Module Overview
  • 6.2 f90_gc
  • 6.3 f90_iostat
  • 6.4 f90_kind
  • 6.5 f90_preconn_io
    • 6.5.1 Procedures
    • 6.5.2 Example
  • 6.6 f90_stat
    • 6.6.1 Parameters
    • 6.6.2 Example
  • 6.7 f90_unix_*
  • 6.8 ieee_*, iso_c_binding, iso_fortran_env
• 7 Modern Fortran API to Posix
  • 7.1 Posix Module Overview
  • 7.2 f90_unix_dir
    • 7.2.1 Parameters
    • 7.2.2 Procedures
  • 7.3 f90_unix_dirent
    • 7.3.1 Procedures
  • 7.4 f90_unix_env
    • 7.4.1 Parameters
    • 7.4.2 Types
    • 7.4.3 Procedures
  • 7.5 f90_unix_errno
    • 7.5.1 Error Handling
    • 7.5.2 Parameters
  • 7.6 f90_unix_file
    • 7.6.1 Parameters
    • 7.6.2 Types
    • 7.6.3 Procedures
  • 7.7 f90_unix_io
    • 7.7.1 Procedures
  • 7.8 f90_unix_proc
    • 7.8.1 Parameters
    • 7.8.2 Procedures
• 8 Standard Fortran 95
  • 8.1 Fortran 95 Program Structure
  • 8.2 Fortran 95 Expressions
  • 8.3 Fortran 95 Statements
  • 8.4 Fortran 95 Intrinsic Procedures
• 9 Fortran 2003 Extensions
  • 9.1 Fortran 2003 Overview
  • 9.2 Object-oriented programming
    • 9.2.1 Type Extension
      • 9.2.1.1 Extending Types [5.0]
      • 9.2.1.2 Polymorphic Variables [5.0]
      • 9.2.1.3 Type Selection [5.0]
      • 9.2.1.4 Unlimited polymorphism [5.2]
      • 9.2.1.5 Ad hoc type comparison [5.3]
    • 9.2.2 Typed allocation [5.1]
    • 9.2.3 Sourced allocation (cloning) [5.1]
    • 9.2.4 Type-bound procedures [5.1]
      • 9.2.4.1 The type-bound procedure part
      • 9.2.4.2 Specific type-bound procedures
      • 9.2.4.3 Generic type-bound procedures
    • 9.2.5 Abstract derived types [5.1]
    • 9.2.6 Object-bound procedures [5.2]
  • 9.3 ALLOCATABLE extensions
    • 9.3.1 Allocatable Dummy Arrays [4.x]
    • 9.3.2 Allocatable Function Results [4.x]
    • 9.3.3 Allocatable Structure Components [4.x]
    • 9.3.4 Allocatable Component Example
    • 9.3.5 The MOVE_ALLOC intrinsic subroutine [5.2]
    • 9.3.6 Allocatable scalars [5.2]
    • 9.3.7 Automatic reallocation [5.2]
  • 9.4 Other data-oriented enhancements
    • 9.4.1 Parameterised derived types [6.0 for kind type parameters, 6.1 for length]
      • 9.4.1.1 Basic Syntax and Semantics
      • 9.4.1.2 More Semantics
      • 9.4.1.3 Assumed type parameters
      • 9.4.1.4 Deferred type parameters
    • 9.4.2 Finalisation [5.3]
    • 9.4.3 The PROTECTED attribute [5.0]
      • 9.4.3.1 Syntax
      • 9.4.3.2 Semantics
      • 9.4.3.3 Example
    • 9.4.4 Pointer enhancements
      • 9.4.4.1 INTENT for pointers [5.1]
      • 9.4.4.2 Pointer bounds specification [5.2]
      • 9.4.4.3 Rank-remapping Pointer Assignment [5.0]
    • 9.4.5 Individual component accessibility [5.1]
    • 9.4.6 Public entities of private type [5.1]
  • 9.5 C interoperability [mostly 5.1]
    • 9.5.1 The ISO_C_BINDING module
      • 9.5.1.1 The kind parameters
      • 9.5.1.2 Using C_PTR and C_FUNPTR
    • 9.5.2 BIND(C) types
    • 9.5.3 BIND(C) variables
    • 9.5.4 BIND(C) procedures
    • 9.5.5 Enumerations
  • 9.6 IEEE arithmetic support [4.x except as otherwise noted]
    • 9.6.1 Introduction
    • 9.6.2 Exception flags, modes and information flow
    • 9.6.3 Procedures in the modules
    • 9.6.4 The IEEE_FEATURES module
    • 9.6.5 IEEE_EXCEPTIONS
      • 9.6.5.1 Types and constants
      • 9.6.5.2 Procedures
    • 9.6.6 IEEE_ARITHMETIC module
      • 9.6.6.1 IEEE datatype selection
      • 9.6.6.2 Enquiry functions
      • 9.6.6.3 Rounding mode
      • 9.6.6.4 Underflow mode
      • 9.6.6.5 Number Classification
      • 9.6.6.6 Test functions
      • 9.6.6.7 Arithmetic functions
  • 9.7 Input/output Features
    • 9.7.1 Stream input/output [5.1]
    • 9.7.2 The BLANK= and PAD= specifiers [5.1]
    • 9.7.3 Decimal Comma [5.1]
    • 9.7.4 The DELIM= specifier [5.1]
    • 9.7.5 The ENCODING= specifier [5.1]
    • 9.7.6 The IOMSG= specifier [5.1]
    • 9.7.7 The IOSTAT= specifier [5.1]
    • 9.7.8 The SIGN= specifier [5.1]
    • 9.7.9 Intrinsic functions for testing IOSTAT= values [5.1]
    • 9.7.10 Input/output of IEEE infinities and NaNs [5.1]
    • 9.7.11 Output of floating-point zero [5.1]
    • 9.7.12 NAMELIST and internal files [5.1]
    • 9.7.13 Variables permitted in NAMELIST
    • 9.7.14 Recursive input/output [5.2]
    • 9.7.15 Asynchronous input/output
      • 9.7.15.1 Basic syntax [5.1]
      • 9.7.15.2 Basic Example
      • 9.7.15.3 The ASYNCHRONOUS attribute [5.2]
      • 9.7.15.4 The WAIT statement [5.2]
      • 9.7.15.5 Execution Semantics
    • 9.7.16 Scale factor followed by repeat count [5.1]
    • 9.7.17 FLUSH statement [5.2]
    • 9.7.18 Defined input/output [6.2]
  • 9.8 Miscellaneous Fortran 2003 Features
    • 9.8.1 Abstract interfaces and the PROCEDURE statement [5.1]
    • 9.8.2 Named procedure pointers [5.2]
    • 9.8.3 Intrinsic modules [4.x]
    • 9.8.4 Renaming user-defined operators on the USE statement [5.2]
    • 9.8.5 The ISO_FORTRAN_ENV module [5.1]
    • 9.8.6 The IMPORT statement [5.1]
    • 9.8.7 Length of names and statements
    • 9.8.8 Array constructor syntax enhancements
    • 9.8.9 Structure constructor syntax enhancements [5.3]
    • 9.8.10 Deferred character length [5.2]
    • 9.8.11 The ERRMSG= specifier [5.1]
    • 9.8.12 Intrinsic functions in constant expressions [5.2 partial; 5.3 complete]
    • 9.8.13 Specification functions can be recursive [6.2]
    • 9.8.14 Access to the command line [5.1]
    • 9.8.15 Access to environment variables [5.1]
    • 9.8.16 Character kind selection [5.1]
    • 9.8.17 Argument passing relaxation [5.1]
    • 9.8.18 The MAXLOC and MINLOC intrinsic functions [5.1]
    • 9.8.19 The VALUE attribute [4.x]
      • 9.8.19.1 Syntax
      • 9.8.19.2 Semantics
      • 9.8.19.3 Example
    • 9.8.20 The VOLATILE attribute [5.0]
    • 9.8.21 Enhanced complex constants [5.2]
    • 9.8.22 The ASSOCIATE construct [5.2]
    • 9.8.23 Binary, octal and hexadecimal constants [5.2]
    • 9.8.24 Character sets [5.1; 5.3]
    • 9.8.25 Intrinsic function changes for 64-bit machines [5.2]
    • 9.8.26 Miscellaneous intrinsic procedure changes [5.2]
• 10 Fortran 2008 Extensions
  • 10.1 Fortran 2008 Overview
  • 10.2 SPMD programming with coarrays [6.2, 7.0]
    • 10.2.1 Overview
    • 10.2.2 Images
    • 10.2.3 Coarrays
    • 10.2.4 Declaring coarrays
    • 10.2.5 Accessing coarrays on other images
    • 10.2.6 Segments and synchronisation
    • 10.2.7 Allocating and deallocating coarrays
    • 10.2.8 Critical constructs
    • 10.2.9 Lock variables
    • 10.2.10 Atomic coarray accessing
    • 10.2.11 Normal termination of execution
    • 10.2.12 Error termination
    • 10.2.13 Fault tolerance
    • 10.2.14 Detailed syntax of coarray features
    • 10.2.15 Intrinsic procedures and coarrays
  • 10.3 Data declaration [mostly 6.0]
  • 10.4 Data usage and computation [mostly 5.3]
  • 10.5 Execution control [mostly 6.0]
  • 10.6 Intrinsic procedures and modules
    • 10.6.1 Additional mathematical intrinsic functions [mostly 5.3.1]
    • 10.6.2 Additional intrinsic functions for bit manipulation [mostly 5.3]
    • 10.6.3 Other new intrinsic procedures [mostly 5.3.1]
    • 10.6.4 Changes to existing intrinsic procedures [mostly 5.3.1]
    • 10.6.5 ISO_C_BINDING additions [6.2]
    • 10.6.6 ISO_FORTRAN_ENV additions
  • 10.7 Input/output extensions [mostly 5.3]
  • 10.8 Programs and procedures [mostly 5.3]
• 11 Appendices
  • 11.1 Mixing Fortran and C
    • 11.1.1 Naming Conventions
    • 11.1.2 Initialisation and Termination
    • 11.1.3 Calling Conventions
    • 11.1.4 Data Types
      • 11.1.4.1 Pointers
      • 11.1.4.2 Derived types
      • 11.1.4.3 Supporting types
    • 11.1.5 SUBROUTINE return types
      • 11.1.5.1 SUBROUTINEs with label arguments
      • 11.1.5.2 SUBROUTINEs with no label arguments
    • 11.1.6 FUNCTION return types
      • 11.1.6.1 Scalar
      • 11.1.6.2 Scalar POINTER functions
      • 11.1.6.3 Array non-POINTER functions
      • 11.1.6.4 Array POINTER functions
    • 11.1.7 Argument types
      • 11.1.7.1 CHARACTER type
      • 11.1.7.2 non-POINTER non-ALLOCATABLE Scalar
      • 11.1.7.3 POINTER and ALLOCATABLE Scalar
      • 11.1.7.4 non-POINTER Array
      • 11.1.7.5 POINTER Array
      • 11.1.7.6 Procedures
      • 11.1.7.7 OPTIONAL arguments
  • 11.2 ASCII Collating Sequence
    • 11.2.1 Printing Characters
    • 11.2.2 Non-printing Characters
• 12 Detailed Contents