DIGITAL Fortran 90
User Manual for
DIGITAL UNIX Systems

M
-m option (cpp)
Machine-code output listing
general description
-machine_code option
Magnetic tape I/O
library routines for #1
library routines for #2
Main program
C language
f90 option for #1
f90 option for #2
statements for
make
options related to
Makefile, example of using
malloc library routine
man command
f90(1)
viewing routine reference pages
Manual optimization
balancing the workload
interchanging loops
MASTER directive #1
MASTER directive #2
MASTER directive #3
-math_library accurate option
effect of -fast option
-math_library fast option
effect on -fpen option
MATMUL intrinsic procedure
selecting version used
Matrix multiplication
selecting MATMUL intrinsic code to use
MAXREC specifier
See also LRM
Memory
allocating and freeing virtual (library routine)
for intermediate storage
Message catalog
location of
transporting
Messages
driver-related errors
issued by compiler
general description
limiting the number
language dialects and standards checking
linking #1
linking #2
run-time format
run-time messages
displaying online explanations
list, numeric order
parallel processing
transporting message file
See also Warning messages
severity
meaning to run-time system
Mixed real/complex operation
-module option
MODULE statement #1
MODULE statement #2
MODULE statement #3
See also LRM
Module variables
accessing in Ladebug
Modules
compiler use of process descriptor limit
declaration example
example compilation
file name suffix
files created and used by compiler #1
files created and used by compiler #2
.mod file suffix
See also LRM
subprograms
to contained subprograms
use association
more command
use with online release notes
-mp option #1
-mp option #2
-mp option #3
mpc_destroy library routine
mpc_in_parallel_region library routine
mpc_maxnumthreads library routine
mpc_my_threadnum library routine
mpc_numthreads library routine
MP_CHUNK_SIZE environment variable
MP_SCHEDTYPE directive #1
MP_SCHEDTYPE directive #2
MP_SCHEDTYPE directive #3
MP_SCHEDTYPE option
for PARALLEL DO directive
for PDO directive #1
for PDO directive #2
for PDO directive #3
MP_SPIN_COUNT environment variable
MP_STACK_SIZE environment variable
MP_THREAD_COUNT environment variable
MP_YIELD_COUNT environment variable
Multiple thread locking routines
N
Name length, symbolic
maximum
NAME specifier
See also LRM
Named common blocks
Privitizing #1
Privitizing #2
Namelist I/O statements
See also LRM
-names keyword options
NaN values (IEEE)
representation in Alpha floating-point data
See also ISNAN intrinsic
Natural boundaries (alignment)
definition of
See also Alignment
Natural storage order
.NE. operator
See also LRM
-nearest_neighbor option
Nesting limits
source code
Nesting of directives
NLSPATH environment variable
use by RTL
nm command
use in locating undefined symbols
NML specifier
See also LRM
-noextend_source option
-nof77 option
-nofor_main option #1
-nofor_main option #2
-nofp_reorder option
effect of -fast option
-noi4 option
-noinclude option #1
-noinclude option #2
-noinclude option #3
-noinclude option #4
-noinline option #1
-noinline option #2
Nonadvancing I/O
-non_shared option #1
-non_shared option #2
-nopad_source option
not function
.NOT. operator
See LRM
NOWAIT clause
for END DO directive
for END SECTIONS
for END SINGLE
NOWAIT option
for END PDO directive #1
for END PDO directive #2
for END PDO directive #3
-nowarn option
-nowsf_main option
-no_archive option (ld)
and creating shared libraries
-no_fp_reorder option
Numeric range
See Data type
Numerical data
output of
O
.o file suffix
-o option #1
-o option #2
and creating shared libraries
effect of -c option
-o output option
-o0 option
-o1 option
-o2 option
-o3 option
-o4 or -o option
-o5 option #1
-o5 option #2
Object file
and cDEC$ directives
contents #1
contents #2
directory used
effect of -gn option on size
effect of -on options on size
effect of optimization level on size #1
effect of optimization level on size #2
effect of optimization level on size #3
linker order of loading
linking
multiple input files and options
naming #1
naming #2
nonshared optimizations
obtaining information about
options controlling size of
passing directly to ld (example)
prevent creation of
recognized file name suffix
renaming
retaining #1
retaining #2
retaining #3
used to create a shared library #1
used to create a shared library #2
used to create a shared library #3
Object library
See Shared library; Archive library
odump command
-om option #1
-om option #2
-omp option #1
-omp option #2
-omp option #3
omp_destroy_lock library routine
OMP_DYNAMIC environment variable
omp_get_dynamic library routine
omp_get_max_threads library routine
omp_get_nested library routine
omp_get_num_procs library routine
omp_get_num_threads library routine
omp_get_thread_num library routine
omp_init_lock library routine
omp_in_parallel library routine
OMP_NESTED environment variable
OMP_NUM_THREADS environment variable #1
OMP_NUM_THREADS environment variable #2
OMP_SCHEDULE environment variable
omp_set_dynamic library routine
omp_set_lock library routine
omp_set_nested library routine
omp_set_num_threads library routine
omp_test_lock library routine
omp_unset_lock library routine
-on (optimization) options #1
-on (optimization) options #2
additional global
automatic inlining
effect of -g options
effect of -gen_feedback option
global
local
loop transformation and software pipelining
-onetrip option
OPEN statement
access mode
ACCESS specifier #1
ACCESS specifier #2
ACTION specifier
APPEND specifier
ASSOCIATEVARIABLE specifier
BLANK specifier
effect of -vms option
BLOCKSIZE specifier #1
BLOCKSIZE specifier #2
BUFFERCOUNT specifier #1
BUFFERCOUNT specifier #2
CARRIAGECONTROL specifier #1
CARRIAGECONTROL specifier #2
effect of -vms option
CONVERT specifier
CONVERT specifier (nonnative data) #1
CONVERT specifier (nonnative data) #2
DEFAULTFILE specifier #1
DEFAULTFILE specifier #2
DEFAULTFILE specifier #3
DEFAULTFILE specifier #4
DEFAULTFILE specifier #5
defaults
converting nonnative data
pathname used #1
pathname used #2
See also LRM
units for preconnected files
DELIM specifier
directory and file name defaults
DISPOSE specifier #1
DISPOSE specifier #2
effect of opening previously open file
effect of -vms option
environment variables
effect of -vms option
ERR specifier
example
example
file organization
importance of specifying
FILE specifier #1
FILE specifier #2
FILE specifier #3
FILE specifier #4
FILE specifier #5
example
file status
FORM specifier #1
FORM specifier #2
I/O statement interdependencies
implied and explicit file open
importance of examining defaults
importance of specifying record type
interdependencies of file and directory
IOSTAT specifier
example
MAXREC specifier
NAME specifier
obtaining file descriptor (library routine)
opening file with user-supplied function (USEROPEN)
ORGANIZATION specifier #1
ORGANIZATION specifier #2
PAD specifier
with fixed-length records
POSITION specifier #1
POSITION specifier #2
READONLY specifier
RECL specifier #1
RECL specifier #2
excluding overhead bytes
obtaining value for unformatted files
option to specify units
performance considerations
specifying for fixed-length records #1
specifying for fixed-length records #2
units
units for unformatted files
record length #1
record length #2
RECORDSIZE specifier
RECORDTYPE specifier #1
RECORDTYPE specifier #2
See also LRM
SHARED specifier #1
SHARED specifier #2
specifiers for efficient I/O #1
specifiers for efficient I/O #2
specifiers identifying
error handling capabilities
file access and position
file and record characteristics
file and unit
file close action
record transfer characteristics
STATUS='SCRATCH' specifier
STATUS specifier #1
STATUS specifier #2
summary of specifiers
TYPE specifier
UNIT specifier #1
UNIT specifier #2
UNIT specifier #3
UNIT specifier #4
USEROPEN specifier #1
USEROPEN specifier #2
using preconnected files
open system call
example USEROPEN function (C)
using to open file
OPENED specifier
See also LRM
OpenMP
conditional compilation
conditional compilation prefixes
fixed source form
free source form
OpenMP parallel directives #1
OpenMP parallel directives #2
ATOMIC #1
ATOMIC #2
ATOMIC #3
BARRIER #1
BARRIER #2
BARRIER #3
CRITICAL #1
CRITICAL #2
CRITICAL #3
CRITICAL #4
CRITICAL #5
DO #1
DO #2
DO #3
DO #4
END CRITICAL #1
END CRITICAL #2
END DO #1
END DO #2
END DO #3
END MASTER #1
END MASTER #2
END MASTER #3
END ORDERED #1
END ORDERED #2
END ORDERED #3
END PARALLEL #1
END PARALLEL #2
END PARALLEL #3
END PARALLEL #4
END PARALLEL DO #1
END PARALLEL DO #2
END PARALLEL SECTIONS #1
END PARALLEL SECTIONS #2
END SECTIONS #1
END SECTIONS #2
END SECTIONS #3
END SINGLE #1
END SINGLE #2
FLUSH #1
FLUSH #2
FLUSH #3
format
MASTER #1
MASTER #2
MASTER #3
ORDERED #1
ORDERED #2
ORDERED #3
PARALLEL #1
PARALLEL #2
PARALLEL #3
PARALLEL #4
PARALLEL #5
PARALLEL DO #1
PARALLEL DO #2
PARALLEL DO #3
PARALLEL SECTIONS #1
PARALLEL SECTIONS #2
PARALLEL SECTIONS #3
prefixes
fixed source form
free source form
SECTION #1
SECTION #2
SECTION #3
SECTIONS #1
SECTIONS #2
SECTIONS #3
SINGLE #1
SINGLE #2
SINGLE #3
summary descriptions
THREADPRIVATE #1
THREADPRIVATE #2
THREADPRIVATE #3
THREADPRIVATE #4
THREADPRIVATE #5
OpenMP run-time library routines
reference page listing
OpenVMS Fortran
and DIGITAL Fortran 90 record types
options for VAX compatibility (f90)
porting code
porting data
Operations
mixed real/complex
Operators
arithmetic
for efficient run-time performance
See also LRM
Optimization #1
Optimization #2
additional global
and performance measurement
automatic inlining
code hoisting
code replication to eliminate branches
common subexpression elimination
compile-time operations
constant pooling
controlling procedure inlining
data-flow and split lifetime analysis
dead code
dead store elimination (unused variables)
effect on
compilation time
debugging
floating-point calculations
for parallel HPF programs
for specific Alpha processor generation #1
for specific Alpha processor generation #2
global
implied DO loop collapsing #1
implied DO loop collapsing #2
inline expansion
inlining procedures #1
inlining procedures #2
instruction scheduling #1
instruction scheduling #2
instruction scheduling #3
instruction scheduling #4
interprocedure analysis
level
summary of options for controlling
limiting loop unrolling #1
limiting loop unrolling #2
linker
local
loop blocking #1
loop blocking #2
loop distribution #1
loop distribution #2
loop fusion #1
loop fusion #2
loop interchange #1
loop interchange #2
loop outer loop unrolling #1
loop outer loop unrolling #2
loop scalar replacement #1
loop scalar replacement #2
loop transformation #1
loop transformation #2
loop transformation #3
loop unrolling
loops #1
loops #2
loops #3
math library use
mixed real/complex operations
multiplication and division expansion
of loops
of multiple source files
effect of -c option
of statement functions and intrinsics
options for
summary
overview of levels
register use
removal optimizations
reordering floating-point operations
setting options with -fast
software pipelining #1
software pipelining #2
source code guidelines for (check list)
speculative execution
summary of levels
summary of -o options
to reduce program size (space optimizations) #1
to reduce program size (space optimizations) #2
-tune keyword
-tune option
using correct options with multiple input files #1
using correct options with multiple input files #2
using dummy aliases
value propagation
OPTIONS statement #1
OPTIONS statement #2
OPTIONS statement #3
See also LRM
specifying unformatted file floating-point format
or function
Order of subscript progression
in I/O
ORDERED clause
for DO directive #1
for DO directive #2
ORDERED directive #1
ORDERED directive #2
ORDERED directive #3
ORDERED option
for PDO directive #1
for PDO directive #2
ORGANIZATION specifier #1
ORGANIZATION specifier #2
See also LRM
otsgetmaxthreads library routine
otsgetnumthreads library routine
otsgetthreadnum library routine
otsinitparallel library routine
otsinparallel library routine
otsstopworkers library routine
Output files
changing output file names (f90)
created by f90
naming (f90)
Output listing #1
Output listing #2
compilation-summary section
machine-code section
options for #1
options for #2
options for #3
Overhead
record
Overriding implicit synchronization