1STATE(1)                         LAM COMMANDS                         STATE(1)
2
3
4

NAME

6       state - Report status of LAM processes.
7

SYNTAX

9       state [-ghklt] <nodes> [<processes>]
10

OPTIONS

12       -g      Print information on system and application processes.
13
14       -h      Print the command help menu.
15
16       -k      Only obtain kernel information on local kernel processes.
17
18       -l      Print supplementary information.
19
20       -t      Print information on system processes only.
21

DESCRIPTION

23       Most  MPI  users  will  probably not need to use the state command; see
24       mpimsg(1) and mpitask(1).  This command is only  installed  if  LAM/MPI
25       was configured with the --with-trillium switch.
26
27       If  no process selection is given on the command line, one line of sta‐
28       tus will be printed for every process on each specified node.
29
30       In default mode, information is organized under the following headings:
31
32       NODE      the nodeid on which the process is running
33
34       INDEX     the LAM process index
35
36       PID       the process identifier from the underlying operating system
37
38       KPRI      the LAM kernel priority
39
40       KSTATE    the LAM kernel execution status:
41
42                 R         running - unencumbered by LAM
43
44                 BR        blocked receiving - The blocked message  event  and
45                           type are also displayed.
46
47                 BS        blocked  sending  -  The  blocked message event and
48                           type are also displayed.
49
50                 F         no longer a LAM process
51
52       PROGRAM   the program filename
53
54       With the -l option, state prints signal information instead  of  execu‐
55       tion status.
56
57       SIGS      all pending but undelivered LAM signals
58
59       Some  event/type  combinations  are  used  by  the  system to implement
60       client/server requests or to implement other message passing libraries.
61       Where possible, state replaces these numbers with the name of the func‐
62       tion that caused the current execution state.  In general, the informa‐
63       tion  provided  by state is low-level and geared toward LAM system pro‐
64       grammers.  MPI programmers are expected to use  mpitask(1)  to  monitor
65       the state of MPI processes.
66
67   Kernel Processes
68       A process must call kenter(2) or be created by loadgo(1) in order to be
69       directly visible to state.  In other cases, limited monitoring is  pos‐
70       sible in one of two ways.  If a process identifier is given, the kernel
71       on each  specified  node  will  be  consulted  for  kernel  information
72       (kstate(2)).  On the local node (use node specifier h) a process can be
73       watched without the assistance of the process management daemon by giv‐
74       ing the process identifier and the -k option.
75

EXAMPLES

77       state n0-7
78           Display  the status of all application processes on nodes 0 through
79           7.
80
81       state n7 i4
82           Display the status of LAM process index 4 on node 7.
83

DIAGNOSTICS

85       If no processes are found, only the title line is displayed.
86

SEE ALSO

88       loadgo(1), mpimsg(1), mpitask(1), doom(1), bfstate(1)
89
90
91
92LAM 7.1.2                         March, 2006                         STATE(1)
Impressum