INTROU(1)

✨ NAME ✨

okay so bestie, welcome to introu — it's the intro to the LAM user interface commands. like, the whole girlies of LAM commands, all in one place. this is your orientation, your vibe check, your little map 💖

🌈 LIST OF COMMANDS 🌈

here's the whole squad, with where they live in the manual and what they do. every single one, no skips, periodt 👑

• bfctl — page bfctl.1 — Control message buffers (optional).
• bfstate — page bfstate.1 — Get message buffer status (optional).
• doom — page doom.1 — Deliver a signal (optional).
• fctl — page fctl.1 — Control remote file access (optional).
• fstate — page fstate.1 — Get remote file status (optional).
• hboot — page hboot.1 — Start LAM on a local node.
• mpicc — page mpicc.1 — Compile LAM C programs.
• mpiCC — page mpicc.1 — Compile LAM C++ programs.
• mpif77 — page mpif77.1 — Compile LAM F77 programs.
• lamboot — page lamboot.1 — Start LAM.
• lamclean — page lamclean.1 — Clean all nodes.
• lamexec — page lamexec.1 — Run non-MPI programs.
• lamgrow — page lamgrow.1 — Add a node.
• lamhalt — page lamhalt.1 — Terminate LAM.
• lamshrink — page lamshrink.1 — Remove a node.
• lamtrace — page lamtrace.1 — Unload trace data.
• loadgo — page loadgo.1 — Execute program (optional).
• mpimsg — page mpimsg.1 — Display MPI buffered messages.
• mpirun — page mpirun.1 — Run an MPI application.
• mpitask — page mpitask.1 — Display MPI processes.
• recon — page recon.1 — Verify LAM setup.
• state — page state.1 — Get process status (optional).
• sweep — page bfctl.1 — Clean out message buffers (optional).
• tkill — page tkill.1 — Terminate local node LAM session.
• tping — page tping.1 — Echo messages to a node.
• lamwipe — page lamwipe.1 — Terminate LAM (deprecated; use lamhalt).

⚠️ heads up bestie — lamwipe is deprecated, so pls use lamhalt instead. it's giving "old money, but like, actually old" 💀

🦋 Node Specification 🦋

you hand the target nodes to the command right there on the command line. nodes get generically called nlist, where list can be just one node identifier or a whole lil group of them. example time:

n1
n1,3,5-10

also, clist is the move when you wanna name specific CPUs on a given machine — and lamboot(1) is what you use to say how many CPUs LAM is allowed to use on each node.

node and CPU identifiers get set up in the boot schema (peep bhost(5)), and you can write them in decimal or hexadecimal notation, either one. and on top of naming nodes explicitly, LAM has special mnemonics for special nodes or a whole group of nodes:

• h — the local node where the command is typed (as in "here") 📍
• o — the origin node where LAM was started with lamboot(1) 🌱
• N — all nodes 🌐
• C — all CPUs 🧠

🌸 Process Specification 🌸

LAM processes? two ways to point at 'em, bestie:

1. by process identifier from the underlying operating system, or
2. by LAM process index.

PIDs are written as plist, where list can be a single PID or a list of PIDs. process indices are written as ilist, where list can be a single index or a list of indices. iykyk 💅

okay that's the whole page!! this was LAM 7.1.4, dated July, 2007 🎀 ty for reading, mwah, go boot something iconic 🫶✨