NAG Fortran Compiler, Release 7.2

NAG^® Fortran Compiler

© 2024 The Numerical Algorithms Group Limited

All rights reserved. No part of this Manual may be reproduced, transcribed, stored in a retrieval system, translated into any language or computer language or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except for the purpose of using the NAG Fortran Compiler.

The copyright owner gives no warranties and makes no representations about the contents of this Manual and specifically disclaims any implied warranties of merchantability or fitness for any purpose.

The copyright owner reserves the right to revise this Manual and to make changes from time to time in its contents without notifying any person of such revisions or changes.

12^th Edition – January 2024

NAG is a registered trademark of:

The Numerical Algorithms Group Limited
The Numerical Algorithms Group Inc
Nihon Numerical Algorithms Group KK

All other trademarks are acknowledged.

NAG Ltd	Nihon NAG KK
30 St Giles'	Hatchobori Frontier Building 2F
OXFORD	4-9-9
OX1 3LE	Hatchobori
United Kingdom	Chuo-ku
Tel: +44 (0)1865 511245 (general)	Tokyo
Tel: +44 (0)1865 311744 (support)	104-0032
Web: www.nag.com	Japan
	Tel: +81 (0)3 5542 6311
	Fax: +81 (0)3 5542 6312
	Web: www.nag-j.co.jp
NAG Inc
801 Warrenville Road, Suite 185
Lisle, IL 60532-4332
USA
Tel: +1 630 971 2337
Fax: +1 630 971 2706
Web: www.nag.com

NAG also has a number of distributors throughout the world. Please contact NAG for further details.

• 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 Fortran 2018 Extensions
  • 11.1 Fortran 2018 Overview
  • 11.2 Data declaration
  • 11.3 Data usage and computation
  • 11.4 Input/output
  • 11.5 Execution control
  • 11.6 Intrinsic procedures and modules
  • 11.7 Program units and procedures
  • 11.8 Advanced C interoperability
  • 11.9 Updated IEEE arithmetic capabilities
  • 11.10 Advanced coarray programming
• 12 Appendices
  • 12.1 Mixing Fortran and C
    • 12.1.1 Naming Conventions
    • 12.1.2 Initialisation and Termination
    • 12.1.3 Calling Conventions
    • 12.1.4 Data Types
      • 12.1.4.1 Pointers
      • 12.1.4.2 Derived types
      • 12.1.4.3 Supporting types
    • 12.1.5 SUBROUTINE return types
      • 12.1.5.1 SUBROUTINEs with label arguments
      • 12.1.5.2 SUBROUTINEs with no label arguments
    • 12.1.6 FUNCTION return types
      • 12.1.6.1 Scalar
      • 12.1.6.2 Scalar POINTER functions
      • 12.1.6.3 Array non-POINTER functions
      • 12.1.6.4 Array POINTER functions
    • 12.1.7 Argument types
      • 12.1.7.1 CHARACTER type
      • 12.1.7.2 non-POINTER non-ALLOCATABLE Scalar
      • 12.1.7.3 POINTER and ALLOCATABLE Scalar
      • 12.1.7.4 non-POINTER Array
      • 12.1.7.5 POINTER Array
      • 12.1.7.6 Procedures
      • 12.1.7.7 OPTIONAL arguments
  • 12.2 ASCII Collating Sequence
    • 12.2.1 Printing Characters
    • 12.2.2 Non-printing Characters
• 13 Detailed Contents