DIGITAL Fortran 90
User Manual for
DIGITAL UNIX Systems

P
-p option
-p option (cpp)
-p0 option #1
-p0 option #2
-p1 and -p option
PAD specifier
See also LRM
Padding source records
-pad_source option
Page fault
and temporary storage
Parallel compiler directives
ATOMIC #1
ATOMIC #2
ATOMIC #3
BARRIER #1
BARRIER #2
BARRIER #3
BARRIER #4
BARRIER #5
BARRIER #6
binding
binding rules checking
CHUNK #1
CHUNK #2
CHUNK #3
COPYIN #1
COPYIN #2
CRITICAL #1
CRITICAL #2
CRITICAL #3
CRITICAL #4
CRITICAL #5
CRITICAL SECTION #1
CRITICAL SECTION #2
CRITICAL SECTION #3
DO #1
DO #2
DO #3
DO #4
DOACROSS #1
DOACROSS #2
DOACROSS #3
END CRITICAL #1
END CRITICAL #2
END CRITICAL SECTION #1
END CRITICAL SECTION #2
END CRITICAL SECTION #3
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 #5
END PARALLEL #6
END PARALLEL DO #1
END PARALLEL DO #2
END PARALLEL DO #3
END PARALLEL DO #4
END PARALLEL SECTIONS #1
END PARALLEL SECTIONS #2
END PARALLEL SECTIONS #3
END PARALLEL SECTIONS #4
END PARALLEL SECTIONS #5
END PDO #1
END PDO #2
END PDO #3
END PSECTIONS #1
END PSECTIONS #2
END PSECTIONS #3
END SECTIONS #1
END SECTIONS #2
END SECTIONS #3
END SINGLE #1
END SINGLE #2
END SINGLE PROCESS #1
END SINGLE PROCESS #2
END SINGLE PROCESS #3
FLUSH #1
FLUSH #2
FLUSH #3
INSTANCE #1
INSTANCE #2
INSTANCE PARALLEL #1
INSTANCE PARALLEL #2
INSTANCE PARALLEL #3
INSTANCE PARALLEL #4
INSTANCE SINGLE #1
INSTANCE SINGLE #2
MASTER #1
MASTER #2
MASTER #3
MP_SCHEDTYPE #1
MP_SCHEDTYPE #2
MP_SCHEDTYPE #3
nesting
ORDERED #1
ORDERED #2
ORDERED #3
PARALLEL #1
PARALLEL #2
PARALLEL #3
PARALLEL #4
PARALLEL #5
PARALLEL #6
PARALLEL #7
PARALLEL DO #1
PARALLEL DO #2
PARALLEL DO #3
PARALLEL DO #4
PARALLEL DO #5
PARALLEL DO #6
PARALLEL SECTIONS #1
PARALLEL SECTIONS #2
PARALLEL SECTIONS #3
PARALLEL SECTIONS #4
PARALLEL SECTIONS #5
PARALLEL SECTIONS #6
PDO #1
PDO #2
PDO #3
PDONE #1
PDONE #2
PDONE #3
PSECTIONS #1
PSECTIONS #2
PSECTIONS #3
SECTION #1
SECTION #2
SECTION #3
SECTION #4
SECTION #5
SECTION #6
SECTIONS #1
SECTIONS #2
SECTIONS #3
SINGLE #1
SINGLE #2
SINGLE #3
SINGLE PROCESS #1
SINGLE PROCESS #2
SINGLE PROCESS #3
summary descriptions #1
summary descriptions #2
TASKCOMMON #1
TASKCOMMON #2
TASKCOMMON #3
TASKCOMMON #4
THREADPRIVATE #1
THREADPRIVATE #2
THREADPRIVATE #3
THREADPRIVATE #4
THREADPRIVATE #5
PARALLEL directive #1
PARALLEL directive #2
PARALLEL directive #3
PARALLEL directive #4
PARALLEL directive #5
PARALLEL directive #6
PARALLEL directive #7
COPYIN clause #1
COPYIN clause #2
COPYIN clause #3
COPYIN option
DEFAULT clause #1
DEFAULT clause #2
DEFAULT clause #3
DEFAULT option
FIRSTPRIVATE clause #1
FIRSTPRIVATE clause #2
FIRSTPRIVATE clause #3
FIRSTPRIVATE option #1
FIRSTPRIVATE option #2
IF clause #1
IF clause #2
LOCAL option
PRIVATE clause #1
PRIVATE clause #2
PRIVATE clause #3
PRIVATE option
REDUCTION clause #1
REDUCTION clause #2
REDUCTION clause #3
SHARED clause #1
SHARED clause #2
SHARED clause #3
SHARED option
PARALLEL DO directive #1
PARALLEL DO directive #2
PARALLEL DO directive #3
PARALLEL DO directive #4
PARALLEL DO directive #5
PARALLEL DO directive #6
COPYIN clause #1
COPYIN clause #2
COPYIN option
DEFAULT clause #1
DEFAULT clause #2
FIRSTPRIVATE clause #1
FIRSTPRIVATE clause #2
FIRSTPRIVATE option
LASTLOCAL option #1
LASTLOCAL option #2
LASTPRIVATE clause #1
LASTPRIVATE clause #2
LOCAL option
MP_SCHEDTYPE option
PRIVATE clause #1
PRIVATE clause #2
PRIVATE option
REDUCTION clause #1
REDUCTION clause #2
REDUCTION option
SHARED clause #1
SHARED clause #2
SHARED option
Parallel execution
DIGITAL Parallel Software Environment #1
DIGITAL Parallel Software Environment #2
directives #1
directives #2
directives #3
directives #4
HPF
HPF run-time messages
KAP for DIGITAL Fortran 90
library routines
options for HPF programs
PSE requirements for multiple compile units
requesting directed decomposition #1
requesting directed decomposition #2
thread model #1
thread model #2
using directed decomposition
Parallel HPF and Non-Parallel HPF Code, calling between
Parallel library routines
OpenMP Fortran API run-time
reference page listing
_Ots and mpc_ threads routines
Parallel region construct #1
Parallel region construct #2
Parallel regions
debugging
PARALLEL SECTIONS directive #1
PARALLEL SECTIONS directive #2
PARALLEL SECTIONS directive #3
PARALLEL SECTIONS directive #4
PARALLEL SECTIONS directive #5
PARALLEL SECTIONS directive #6
COPYIN clause #1
COPYIN clause #2
COPYIN option
DEFAULT clause #1
DEFAULT clause #2
FIRSTPRIVATE clause #1
FIRSTPRIVATE clause #2
FIRSTPRIVATE option
LASTPRIVATE clause #1
LASTPRIVATE clause #2
LOCAL option
PRIVATE clause #1
PRIVATE clause #2
PRIVATE option
REDUCTION clause #1
REDUCTION clause #2
SHARED clause #1
SHARED clause #2
Parallel Software Environment #1
Parallel Software Environment #2
See also Parallel execution; -wsf option
Parallel threads routines
PARAMETER statements, alternative syntax for
Parentheses in expressions
maximum allowed
Pathname
absolute
application of defaults
directory
effect of DEFAULTFILE specifier
I/O statements default use #1
I/O statements default use #2
in I/O statements
OPEN statement specifiers #1
OPEN statement specifiers #2
relative
setting environment variables for
tilde character (~) #1
tilde character (~) #2
tilde character (~) #3
PDO directive #1
PDO directive #2
PDO directive #3
BLOCKED option #1
BLOCKED option #2
CHUNK option #1
CHUNK option #2
FIRSTPRIVATE option #1
FIRSTPRIVATE option #2
LASTLOCAL option #1
LASTLOCAL option #2
LOCAL option #1
LOCAL option #2
MP_SCHEDTYPE option #1
MP_SCHEDTYPE option #2
MP_SCHEDTYPE option #3
ORDERED option #1
ORDERED option #2
PRIVATE option #1
PRIVATE option #2
REDUCTION option #1
REDUCTION option #2
PDONE directive #1
PDONE directive #2
PDONE directive #3
Performance
arithmetic operators and run-time performance
array use efficiency #1
array use efficiency #2
-assume dummy_aliases option
checking
for unaligned data
process limits #1
process limits #2
choosing optimal record type
compilation #1
compilation #2
compilation times and optimization levels
controlling procedure inlining
cord and feedback files
data alignment efficiency
data types and run-time performance #1
data types and run-time performance #2
DO loop coding and run-time performance
effect of formatted files on run-time performance
equivalence and run-time performance
feedback files
-fp_reorder option
internal subprograms and run-time performance
limiting loop unrolling
measuring optimized programs
measuring using shell commands
mixed data type operations and run-time performance
nonshared object optimizations #1
nonshared object optimizations #2
OPEN statement specifiers for efficient I/O
optimization for Alpha processor generation #1
optimization for Alpha processor generation #2
optimization levels
options
controlling alignment
controlling dummy aliases
controlling floating-point calculations
controlling parallel execution #1
controlling parallel execution #2
controlling parallel execution #3
for run-time efficiency
nonshared object optimizations
related to math library use
related to profiling #1
related to profiling #2
related to profiling #3
that improve performance
that slow performance
profiling code
basic block sampling
call graph information
PC sampling
source line sampling
realistic measurements
record I/O buffers
redirecting scrolled output and run-time performance
reordering floating-point operations
run-time
I/O efficiency
source code guidelines for run-time efficiency
statement functions and run-time performance
timing
using routines #1
using routines #2
using routines #3
using shell commands
perror library routine #1
perror library routine #2
-pg option
physical_to_abstract library routine
-pipeline option
pixie command
for basic block sampling
for source line sampling
related f90 options
use with f90
use with feedback files
Pointers
C language
effect of explicit interface
passed between DIGITAL Fortran 90 and C #1
passed between DIGITAL Fortran 90 and C #2
passing as arguments
passing between C and DIGITAL Fortran 90
See also LRM
use with DIGITAL Fortran 77
POSITION specifier #1
POSITION specifier #2
See also LRM
-pprof option #1
-pprof option #2
Preconnected files
OPEN statement
Prefixes
DIGITAL Fortran parallel directives
OpenMP conditional compilation
fixed source form
free source form
OpenMP directives
Preprocessor
improving run-time performance
See cpp preprocessor
Preprocessor symbols
defining
undefining
PRINT statement
See also LRM
printenv command (shell)
viewing environment variables
PRIVATE clause
for DO directive #1
for DO directive #2
for DO directive #3
for PARALLEL directive #1
for PARALLEL directive #2
for PARALLEL directive #3
for PARALLEL DO directive #1
for PARALLEL DO directive #2
for PARALLEL SECTIONS directive #1
for PARALLEL SECTIONS directive #2
for SECTIONS directive #1
for SECTIONS directive #2
for SECTIONS directive #3
for SINGLE directive #1
for SINGLE directive #2
for SINGLE directive #3
using to resolve dependences
PRIVATE option
for PARALLEL directive
for PARALLEL DO directive
for PARALLEL SECTIONS directive
for PDO directive #1
for PDO directive #2
for PSECTIONS directive
for SINGLE PROCESS directive
Privitizing named common blocks #1
Privitizing named common blocks #2
Procedure interface
argument passing rules
array arguments
C language main program
-nofor_main option
changing default passing mechanisms #1
changing default passing mechanisms #2
DIGITAL Fortran 77 and DIGITAL Fortran 90
passing arguments
DIGITAL Fortran 90 and C #1
DIGITAL Fortran 90 and C #2
DIGITAL Fortran 90 and C #3
DIGITAL Fortran 90 descriptor format
DIGITAL Fortran 90 subprograms
explicit interface
implicit interface
modules #1
modules #2
pointer arguments
procedure interface block #1
procedure interface block #2
See also LRM
types of subprograms
Procedures
analyzing performance
inlining #1
inlining #2
inlining #3
inlining #4
inlining #5
inlining #6
compiling multiple files
See also LRM
types of subprograms
Process control
library routines #1
library routines #2
Process information
returning (library routine) #1
returning (library routine) #2
returning (library routine) #3
returning (library routine) #4
Process limits
checking
increasing file descriptor limit
increasing stack size
prof command
for basic block sampling
for PC sampling
for source line sampling
options to limit report contents
related f90 options
use with f90
use with feedback files
PROFDIR
environment variable
Profiling code
basic block sampling
call graph information
DIGITAL UNIX tools
options related to #1
options related to #2
parallel
PC sampling
source line sampling
timing program execution
using routines #1
using routines #2
using shell commands
with f90
Program
compiling #1
compiling #2
compiling #3
compiling #4
parallel execution
HPF
KAP preprocessor
related HPF command-line options
running
section
See also Executable program
size
dividing large source programs (fsplit)
effect of -call_shared option
effect of optimization levels
insufficient virtual memory run-time error
limitations
process stack size
space optimizations
terminating (library routine)
threaded execution
related command-line options #1
related command-line options #2
related command-line options #3
transportability
See Compatibility
units
creating modules for
PROGRAM statement #1
PROGRAM statement #2
See also LRM
Program transportability
PSECTIONS directive #1
PSECTIONS directive #2
PSECTIONS directive #3
FIRSTPRIVATE option
LOCAL option
PRIVATE option
-pthread option
putc library routine