Contents

1. Introduction
2. About Sun Studio 11 Fortran 95
3. New and Changed Features
4. Software Corrections
5. Problems and Workarounds
6. Limitations and Incompatibilities
7. Documentation Errors
8. Required Patches for Fortran 95
9. Shippable Libraries

---

A. Introduction

This document contains information about this release of the Sun[tm] Studio 11 Fortran 95 compiler, f95. This document describes the software corrections, known problems, limitations, and incompatibilities of this release.

Product Documentation

Product documentation for this release of Sun Studio includes the following:

• Release Notes: Available on the Sun Studio web site at http://developers.sun.com/sunstudio/documentation/ss11/release_notes.html. Information in the release notes updates and extends information in all readme files.
• Compilers and Tools Documentation: Product man pages, HTML versions of readmes, and manuals can be accessed from /installation_directory/SUNWspro/docs/index.html. The default installation directory for Solaris platforms is /opt. If your Sun Studio  software is not installed in the default /opt directory, ask your system administrator for the equivalent path on your system. You can access the following documents from the index.html page:
• Fortran User's Guide
• Fortran Programming Guide
• OpenMP API User's Guide
• Math Libraries Readme
• Numerical Computation Guide
• Incremental Link Editor Readme
• Sun Performance Library Readme
• Sun Performance Library Reference Manual
• Sun Performance Library User's Guide for Fortran and C
• IDE Documentation: Online help for all components of the Sun Studio IDE can be accessed from the Help menu in the IDE.
• Developer Resources Portal: For technical articles, code samples, documentation, and a knowledge base, see the developers portal at http://developers.sun.com/prodtech/cc.  For additional information, see the Support page on the portal.

Note - If your compiler collection software has not been installed in the default /opt directory, ask your system administrator for the equivalent path on your system.

---

B. About Sun Studio 11 Fortran 95

This readme has been updated with information describing new and changed features in this release of Version 8.2 of f95.

Version 8.2 of the Fortran 95 compiler f95 is released as a component of Sun Studio 11 for Solaris[tm] platforms, and is available on Solaris Operating System (Solaris OS) versions 8, 9, and 10, on SPARC(R)  and x86 platforms.

The previous version of the compiler was version 8.1, released with Sun Studio 10 .

---

C. New and Changed Features

This section describes new and changed features for this release of Fortran 95. For details on any of the compiler options, see the Fortran User's Guide and the f95(1) man page.

Changes introduced in the Sun Studio 11 Fortran 95 Compiler:

• New -xmodel Option

  The new -xmodel option lets you specify the kernel, small, or medium memory models on the 64-bit AMD architecture. If the size of your global and static variables exceeds two gigabytes, specify -xmodel=medium. Otherwise, don't change the default -xmodel=small setting.

• New Flags For The -xvector Option on x86 SSE2 Platforms

  The -xvector option enables automatic generation of calls to the vector library functions and/or the generation of the SIMD (Single Instruction Multiple Data) instructions. This option now offers the following expanded syntax on x86 SSE2 platforms:
  
  -xvector [=[ {yes|no} ] [ [ no%]lib,[no%]simd, %none] ]

  On SPARC platforms, only -xvector=yes or -xvector=no are enabled.
  
  You must use default rounding mode -fround=nearest when you use this option.

  -xvector=[no%]lib does [not]enable the compiler to transform math library calls within loops into single calls to the equivalent vector math routines when such transformations are possible. This could result in a performance improvement for loops with large loop counts.

  -xvector=[no%]simd does [not]direct the compiler to use the native x86 SSE SIMD instructions to improve performance of certain loops. The compiler can only accept this switch if the target architecture supports SIMD instructions. For example, you must specify -xarch=amd64 or -xarch=generic64. You must also specify an optimization level of -xO3 or above as well as -xdepend with -xvector=simd.

  You will get better performance if you specify both -xvector=simd and -fsimple=2 than with -xvector=simd alone. However, your floating point results can be slightly different because -fsimple=2 allows reordering of floating-point operations.

  The default is -xvector=%none. If you specify -xvector, but do not provide a flag, the compiler assumes -xvector=lib.

  The compiler includes the libmvec libraries in the load step. If you specify -xvector=lib at compile time, you must also specify it at link time.

  This option overrides previous instances so -xvector=%none overrides a previously specified -xvector=lib.

  The -xvector=yes option may be deprecated in a future release. Specify -xvector=lib instead.

  The -xvector=no option may be deprecated in a future release. Specify -xvector=no%lib,no%simd instead.

• STACKSIZE Environment Variable Enhancement:
  The syntax of the STACKSIZE environment variable has been enhanced to include a units keyword. Prior to this enhancement, the STACKSIZE environment variable accepted an integer that denoted the slave thread stacksize in Kilobytes.
  For example: % setenv STACKSIZE 8192 sets the slave thread stack size to 8 MB.
  WIth this release, the STACKSIZE environment variable now accepts a integer value with a qualifying letter that specifies whether the size is in Bytes, Kilobytes, Megabytes, or Gigabytes:
      1235B for 1235 Bytes, 1235K for 1235 Kilobytes, 1235M for 1235 Megabytes, 1235G for 1235 Gigabytes.
  The default for an integer value without a suffix letter is still Kilobytes.
  
  
• -xpagesize Options Available on x86 Platforms:
  The -xpagesize, -xpagesize_heap, -xpagesize_stack compiler options for cc, CC, and f95 are now enabled for x86 platforms as well as SPARC.
      
• New UltraSPARC T1 and UltraSPARC IVplus Targets:
  Sun Studio compilers generate code for these new SPARC hardware processors. Compile with -xtarget=ultraT1 or -xtarget=ultra4plus for these processors.

For information about other Sun Studio components, see the What's New manual. You can find this manual as part of the documentation installed with the software, at file:/opt/SUNWspro/docs/index.html . You can also find the What's New manual in the Sun Studio 11 collection at http://developers.sun.com/sunstudio/documentation/ss11/index.html

---

D. Software Corrections

No information at this time.

---

E. Problems and Workarounds

This section discusses known software problems in this release and possible workarounds for those problems.

• Optimization Levels and -xipo
On x86 platforms, when compiling and linking with -xipo, all parts of the program must be compiled with the same optimization level.
    
  
• OpenMP:  INTERVAL Reduction Variables
  Reduction clauses with INTERVAL type variables on SPARC platforms are not supported, and will give incorrect results.

No additional information.  Check the Support page on the Compilers and Tools portal, http://developers.sun.com/sunstudio/support/ for latest information.

---

F. Limitations and Incompatibilities

This section discusses limitations and incompatibilities with systems or other software.

• Format of module information files (.mod)

The format of a module information file (.mod file) changed in recent releases of the compiler. The compiler will process .mod files generated by previous releases without any problems, except for one case:

 MODULE NEW
 USE OLD
 END MODULE

If module OLD was compiled by a previous release of the f95 compiler and module NEW is compiled by the current release, there could be problems. To avoid these problems, recompile the older module, OLD, with this release of the compiler.

Each compiled .mod file has a version number associated with the compiler that generated it. Module files compiled by this release of the compiler have "Version 1.0". You can determine the version number of a compiled .mod file by using the fdumpmod(1) utility.

Previous releases of the f95 compiler introduced incompatibilities that carry forward to this release of the compiler and should be noted if you are updating from earlier f95 releases. The following incompatibilies are worth noting:

• Interval Function Calls

The interface (ABI) for interval function calls was changed in the Forte Developer 7 release, and is not binary compatible with the ABI for Sun WorkShop 6 update 2 and earlier releases. This means that programs and libraries with interval functions must be recompiled with the current f95. In particular, Fortran interval functions called by C++, and C++ interval functions called by Fortran, must be recompiled with the f95 and CC compilers.

• CHARACTER*1 Call-by-Value Interfaces

Fortran 95 version 7.0 introduced an incompatibility with the previous release in the way it passes CHARACTER*1 in subprogram calls designated with a VALUE attribute in interface blocks.

The difference is transparent for correctly written Fortran 95 programs compiled with versions 7.0 or 8.0 f95. This difference becomes an issue when the caller and the called routine are compiled with different compiler releases, or when one or the other is written in C.

Consider this example:

main.f:
        interface
        subroutine s(c1,c2)
        character*1, value :: c1   !  c1 is call-by-value
        character*2 :: c2    
        integer I
        end subroutine
        end interface
        call s('A', 'BC')
        end

sub.f:
        subroutine s(c1,c2)
        character*1, value :: c1
        character*(*) :: c2            
        print *, 'You passed ',c1,' and ',c2
        END

With main.f compiled with the previous release of the compiler and sub.f compiled with this release, the program incorrectly produces:

 You passed A and B

The compiler now passes call-by-value CHARACTER *1 data without a character length parameter. In all other cases, character data is passed with a hidden length parameter generated automatically by the compiler as part of the call. (Note that Fortran 95 restricts call-by-value of CHARACTER data to only CHARACTER*1 .)

Similarly, C routines calling or called by Fortran 95 routines that expect call-by-value CHARACTER*1 data must be rewritten to no longer expect a length parameter for those arguments.

• Binary Incompatibility with Forte Developer 6 update 2 Object Files

Version 7.0 f95 fixed a problem where the -aligncommon flag did not properly align derived types or variables typed with an explicit KIND.

For example, the following program:

    character*26 type

    logical(1) :: l1 = .true.
    real(16)   :: r16 = 1
    integer(2) :: i2 = 1

    common /A/ l1,r16,i2

    print *, '--------------------------------------------'
    print *, ' Fortran 95 COMMON-block alignment in Bytes '
    print *, '--------------------------------------------'

    type = 'logical(1)'
    print *, type, loc(l1)-loc(l1)

    type = 'real(16)'
    print *, type, loc(r16)-loc(l1)

    type = 'integer(2)'
    print *, type, loc(i2)-loc(l1)

    end

when compiled with FD6u2 Fortran 95 (version 6.2) with -aligncommon produces the following output:

 --------------------------------------------
  Fortran 95 COMMON-block alignment in Bytes 
 --------------------------------------------
 logical(1)                 0
 real(16)                   8
 integer(2)                 24

whereas it should produce (and does, with versions 8.0 and 7.0 f95 compiler)

 --------------------------------------------
  Fortran 95 COMMON-block alignment in Bytes 
 --------------------------------------------
 logical(1)                 0
 real(16)                   1
 integer(2)                 17

If you compile with the -aligncommon option you should not mix FD6U2 object files with objects produced by the version 8.0 or 7.0 compiler if the common blocks contain either derived types or variables typed with an explicit kind value.

The following general items should also be noted:

Linking on SPARC V9 Platforms under OS releases since Solaris 7:

Many static system libraries, such as libm.a and libc.a, are not available in recent Solaris OS releases on SPARC V9 platforms. Only dynamic, shared libraries, libm.so and libc.so, are provided. This means that -Bstatic and -dn compiler options could cause linking errors on SPARC V9 platforms. Applications must use dynamic libraries in these cases.

To explicitly link with a static version of a user library while linking dynamically system libraries, use a command line that looks something like:

  f95 -o prog prog.f -Bstatic -lxyz -labc -Bdynamic

This will link with libxyz.a and libabc.a , but all other libraries, including system libraries, will be linked dynamically.

Array Intrinsic Functions Use Global Registers:

The array intrinsic functions ANY, ALL, COUNT, MAXVAL, MINVAL, SUM, PRODUCT, DOT_PRODUCT, and MATMUL are highly tuned for the appropriate SPARC platform architectures. As a result, they use the global registers %g2, %g3, and %g4 as scratch registers.

User code should not assume these registers are available for temporary storage if calls are made to the array intrinsics listed above. Data in these registers will be overwritten when the array intrinsics are called.

F95 Modules in Archive Libraries Not Included In Executable:

The debugger dbx requires all object files used in the compilation  to be included in the executable file.  Usually, programs satisfy this requirement with no extra work on the part of the user.  An exceptional case arises from the use of archives containing modules. If a program uses a module, but does not reference any of the procedures or variables in the module, the resulting object file will not contain references to the symbols defined in the module. The linker only links with a object file from an archive if there is a reference to a symbol defined in the object file.  If there is no such reference, the object file will not be included in the executable file.  Dbx will give a warning when it tries to find the debugging information associated with the module that was used. It will not be able to provide information about the symbols whose debugging information is missing.

Use the -u linker option to work around this problem. This option takes a symbol as its option argument.  It adds that symbol to the set of undefined linker symbols, so it will have to be resolved.  The linker symbol associated with a module is normally the module name with all letters in lower case followed by an underscore.

For example, to force the object file containing the module MODULE_1 to be taken from an archive, specify the linker option -u module_1_.  If  linking using the f95 command, use the  -Qoption ld -umodule_1_ on the command line.

Fortran 95 Derived Types and SCCS:

Because of SCCS's use of the percent sign (%) to delimit single-letter ID keywords, users with Fortran 95 programs that define structure components with single-letter names could see unexpected results when maintaining their source code files under SCCS.

This is not necessarily a bug in the compiler, but rather a conflict between the Fortran 95 language and the SCCS source code control system.

It can be avoided by not using single letter names for structure components if you plan to manage your source codes with SCCS, or by retrieving the files with the SCCS get option -k which ignores these keywords, or by placing blanks around the %, as in X % Y % Z .

---

G. Documentation Errors

No information at this time.  Additional information may be made available at http://developers.sun.com/sunstudio/

---

H. Required Patches for Fortran 95

For information about required and optional patches for this release, see the Sun Studio release notes at http://developers.sun.comprodtech/cc/documentation/ss11/release_notes.html.

---

I. Shippable Libraries

If your executable uses a Sun dynamic library listed at http://developers.sun.com/sunstudio/documentation/ss11/mr/runtime.libraries.html, your license includes the right to redistribute the library to your customer.

You cannot redistribute or otherwise disclose the header files, source code, object modules, or static libraries of object modules in any form.

The License to Use appears in the End User Object Code License, which you can view from the back of the plastic case containing the CD-ROM.

---

---

Copyright © 2005 Sun Microsystems, Inc. All rights reserved. Use is subject to license terms.