Question 1

How do I stop several Claude Code or Codex agents from fighting over my GPU?

Accepted Answer

Install FIREFIGHTER. New sessions read the rule block and register the MCP server, so they queue GPU work through firefighter run or firefighter_submit and wait their turn.

Question 2

Why do I get CUDA out of memory when two AI agents train at once?

Accepted Answer

Each agent sees the whole card free and starts its own job, so together they ask for more VRAM than the card has. FIREFIGHTER starts a GPU job only when its VRAM need fits on one card next to what's already running and your reserve; the need is learned from that command's past runs.

Question 3

Can several AI agents share one GPU?

Accepted Answer

Yes. They take turns, or run side by side when the card has room (gpu.max_jobs per card). With more than one GPU, each card takes jobs on its own and a job goes to the least busy card that fits.

Question 4

Is there an MCP server for queuing GPU jobs?

Accepted Answer

firefighter mcp, registered for every agent setup finds. Tools: firefighter_submit, firefighter_wait, firefighter_status, firefighter_job, firefighter_plan.

Question 5

Will it slow down my games?

Accepted Answer

It's designed to do the opposite. Fullscreen apps and 3D-engine use switch you to the "game" state (on macOS and Linux, list your games in reserve.exclusive_apps). Running GPU jobs get a cooperative pause; jobs that ignore it are suspended. If your game then needs the memory a job holds, that job checkpoints (or, if it ignored the pause, is stopped) and requeues. No new GPU jobs start until you're done.

Question 6

Does it work with Docker and Kubernetes?

Accepted Answer

Containers you run yourself count as your load and are protected by the reserve. firefighter run -- docker run ... labels the container so FIREFIGHTER can pin, pause and stop it like any other job.

Question 7

What if an agent ignores the rules?

Accepted Answer

If an agent CLI starts something that stays heavy for 60 seconds outside FIREFIGHTER, it gets adopted. It's put on idle priority, charged to its project, and evicted first under pressure.

Question 8

Can agents cheat the market?

Accepted Answer

Prices are posted, not auctioned, so there's no bidding game. A bid of 10 costs double and buys at most 30 minutes of lead. Budgets belong to projects, so running more jobs doesn't buy more turns. Runtimes are checked against each command's history. Agents can read the policy but not change it unless you set guard.agents_can_configure. An agent with shell access could still edit FIREFIGHTER's files directly. FIREFIGHTER is a referee for cooperative agents, not a security boundary.

Question 9

Does it need an NVIDIA GPU?

Accepted Answer

No. It reads NVIDIA (nvidia-smi plus Windows GPU counters) and AMD (rocm-smi). On Apple silicon it schedules unified memory as RAM. Without a GPU it still schedules CPU and RAM.

Your AI agents won't set
your GPU on fire.

Turns, not fights

A reserve that follows you

Least-served goes first

Posted prices, one line

Skip the line, pay for it

Polite, then firm

Every agent knows

Your AI agents won't setyour GPU on fire.