1BHOST(5) LAM FILE FORMATS BHOST(5)
2
6 bhost - LAM boot schema (host file) format
7
9 #
10 # comments
11 #
12 <machine> [cpu=<cpucount>] [user=<userid>]
13 <machine> [cpu=<cpucount>] [user=<userid>]
14 ...
15
17 A boot schema describes the machines that will combine to form a multi‐
18 computer running LAM. It is used by recon(1) to verify initial condi‐
19 tions for running LAM, by lamboot(1) to start LAM, and by lamhalt(1) to
20 terminate LAM (note that lamwipe(1) has been deprecated by the
21 lamhalt(1) command).
22 The particular syntax of a LAM boot schema is sometimes called the
24 "host file" syntax. It is line oriented. One line indicates the name
25 of a machine, typically the full Internet domain name, an optional num‐
26 ber of CPUs available on that machine, and optionally the userid with
27 which to access it.
28 Common boot schema for a particular site may be created by the system
30 administrator and placed in the installation directory under etc/.
31 They typically start with the prefix bhost. Individual users usually
32 create their own boot schema, especially if the configurations are sim‐
33 ple.
34
36 Note that lamboot resolves all names listed in bhost on the node in
37 which lamboot was invoked on. The lamboot(1) man page contains infor‐
38 mation about address resolution, examples on how to handle multiple
39 network interface cards (NICs) in a node, etc.
40
42 Here is an example three node boot schema:
43 #
45 # example LAM host file
46 #
47 server.cluster.example.com schedule=no
48 beowulf1.cluster.example.com cpu=2
49 beowulf2.cluster.example.com
50 beowulf2.cluster.example.com
51 somewhere.else.example.com user=guest
52 Note that the "guest" ID is significant, since the user has an alter‐
54 nate login ID on somewhere.else.example.com. Additionally note that
55 beowulf1 has a CPU count of 2 listed (a CPU count of 1 is assumed if it
56 is not given). This value is used by mpirun(1), MPI_Comm_spawn(2), and
57 MPI_Comm_spawn_multiple(2) for the "C" (or CPU) notation that specifies
58 how many ranks to start. This is particularly useful for running on
59 SMP machines.
60 Note the schedule=no clause. This means that LAM will boot a daemon on
62 that node, but by default, will not launch any MPI processes on that
63 node. This is handy for when you want to control your MPI applications
64 from one node (e.g., a server), but don't want to run any MPI applica‐
65 tions on it. In some environments this is the default (e.g., BProc).
66 See the LAM User's Guide for more details.
67 beowulf2 is listed twice, but has no specific CPU count listed. In
69 this case, LAM will keep a running tally of the total number of CPUs
70 for that host. Hence, LAM will calculate that beowulf2 has two CPUs
71 available for use. Calculating the number of CPUs by counting occu‐
72 rances of a hostname is useful in a batch environment where a hostfile
73 may list the same hostname multiple times, indicating that the batch
74 scheduler has allocated multiple CPUs for a single job (e.g., PBS oper‐
75 ates this way).
76 For the above-mentioned schema, the command "mpirun C foo" would start
78 five instances of the foo program; two on beowulf1, two on beowulf2,
79 and one on somewhere.else.
80
82 $LAMHOME/etc/bhost.def default boot schema file
83
85 LAM User's Guide, lamboot(1), lamhalt(1), mpirun(1), MPI_Comm_spawn(1),
86 MPI_Comm_spawn_multiple(1), recon(1), lamwipe(1)
87LAM 7.1.2 March, 2006 BHOST(5)