introf(2) - Linux man page
Name
INTROF - introduction to LAM Fortran programming routines
List Of Functions
Name Appears Description
- on Page
*FTTIME
FTTIME.3
Get timer value in seconds.
#F4LTOT
TTOL.3
Convert REAL*4 (local to LAM).
#F4MLTT
TTOL.3
Convert REAL*4 array (local to LAM).
#F4MTTL
TTOL.3
Convert REAL*4 array (LAM to local).
#F4TTOL
TTOL.3
Convert REAL*4 (LAM to local).
#F8LTOT
TTOL.3
Convert REAL*8 (local to LAM).
#F8MLTT
TTOL.3
Convert REAL*8 array (local to LAM).
#F8MTTL
TTOL.3
Convert REAL*8 array (LAM to local).
#F8TTOL
TTOL.3
Convert REAL*8 (LAM to local).
#GALL
GNODES.3
Get all nodeids.
#GCASTS
GNODES.3
Get all castids.
#GCOMPS
GNODES.3
Get all compute nodeids.
#GJONES
GNODES.3
Get all neighbouring nodeids.
#GNODES
GNODES.3
Get array of type matching nodeids.
#GOTBS
GNODES.3
Get all OTB nodeids.
*IGNALL
IGNTP.2
Get total number of nodes.
*IGNCMP
IGNTP.2
Get number of compute nodes.
*IGNDID
IGNTP.2
Get local node ID.
*IGNDTP
IGNTP.2
Get local node type.
*IGNJON
IGNTP.2
Get number of neighbouring nodes.
*IGNOTB
IGNTP.2
Get number of OTB nodes.
*IGNTP
IGNTP.2
Get number of type matching nodes.
*IGORGN
IGNTP.2
Get origin node ID.
*IGRTP
IGNTP.2
Get remote node type.
#I2LTOT
TTOL.3
Convert INTEGER*2 (local to LAM).
#I2MLTT
TTOL.3
Convert INTEGER*2 array (local to LAM).
#I2MTTL
TTOL.3
Convert INTEGER*2 array (LAM to local).
#I2TTOL
TTOL.3
Convert INTEGER*2 (LAM to local).
#I4LTOT
TTOL.3
Convert INTEGER*4 (local to LAM).
#I4MLTT
TTOL.3
Convert INTEGER*4 array (local to LAM).
#I4MTTL
TTOL.3
Convert INTEGER*4 array (LAM to local).
#I4TTOL
TTOL.3
Convert INTEGER*4 (LAM to local).
#KENTR
KENTR.2
Register self process.
#KXIT
KXIT.2
Terminate self with cleanup.
#NPROB
NPROB.2
Probe receivable messages.
#NRCV
NSND.2
Receive network message.
#NSND
NSND.2
Send network message.
#PRCV
PSND.2
Receive message (physical layer).
#PRCVC
PSND.2
Close receiver physical channel.
#PRCVO
PSND.2
Open receiver physical channel.
#PSND
PSND.2
Send message (physical layer).
#PSNDC
PSND.2
Close sender physical channel.
#PSNDO
PSND.2
Open sender physical channel.
#RPLDGO
RPLDGO.2
Load and run a program on a node.
#RPWT
RPWT.2
Wait for a child process to exit.
#lamf_rfclose
lamf_rfopen.2
Close file.
#lamf_rfopen
lamf_rfopen.2
Open file.
#lamf_rfread
lamf_rfread.2
Read file.
#TRCV
TSND.2
Receive message (transport layer).
#TRROR
TRROR.3
Print most recent error message.
#TSND
TSND.2
Send message (transport layer).
#lamf_rfwrite
lamf_rfread.2
Write file.
#MPI_SPAWN
MPI_Spawn.2
Create processes.
- LAM/MPI Extensions
- Name
Appears
Description
on Page
#MPIL_COMM_GPS
MPIL_Comm_id.2
Get LAM coordinates for MPI process.
#MPIL_COMM_ID
MPIL_Comm_id.2
Get communicator ID.
#MPIL_COMM_PARENT
MPIL_Spawn.2
Get parent intercommunicator.
#MPIL_SIGNAL
MPIL_Signal.2
Deliver a signal.
#MPIL_SPAWN
MPIL_Spawn.2
Create processes.
#MPIL_TRACE_ON
MPIL_Trace_on.2
Enable trace collection.
#MPIL_TRACE_OFF
MPIL_Trace_on.2
Disable trace collection.
#MPIL_TYPE_ID
MPIL_Comm_id.2
Get datatype ID.
#MPIL_UNIVERSE_SIZE
MPIL_Spawn.2
Get number of nodes.
- LAM/MPI Extensions
Notes
All names marked by a "#" are subroutine calls and have as their last parameter ierror which returns the error code in case of an unsuccessful call. It is therefore recommended to check for the value of ierror immediately after subroutine calls. All names marked by a "*" are function calls and return some useful value other than ierror.
Simple rules for creating FORTRAN subroutine/function names from C functions are listed below.
Mappings from substrings in C functions onto shorter substrings or single letters in FORTRAN wrappers:
C --> FORTRAN
- TRY
T
GET
G
NODE
ND
TYPE
TP
- All functions that return an integer or short integer are prefixed with the letter "I". All functions that return a single precision or double precision floating point value are prefixed with the letter "F". This is to minimize the necessity of declaring functions. However functions that return a short integer or a double precision floating point number still have to be declared.
If after applying the above two rules, the FORTRAN name still exceeds six characters in length, it is compressed to six by dropping the least essential vowels.
- All functions that return an integer or short integer are prefixed with the letter "I". All functions that return a single precision or double precision floating point value are prefixed with the letter "F". This is to minimize the necessity of declaring functions. However functions that return a short integer or a double precision floating point number still have to be declared.
