Models, artifacts, and runtimes · Lesson 11 of 16 · 7 min

Inspect accelerator availability

Accelerator availability has several layers: device visibility, driver/runtime support and free memory. A visible GPU does not prove that the model runtime can use it.

Open the lab Course map

Mastery 0/3Run itExplain itApply it

learner@clairos:~ $

↑ history · Tab complete · Ctrl+C cancel · Esc, Tab to leaveExit 0

Your objectives

Hints are always available. Wrong turns never cost XP.

1
Read the case evidence
Read /opt/ai-lab/cases/aiops-models-runtimes-011.json. Separate the recorded facts from your assumptions.
Hint & worked example
Accelerator availability has several layers: device visibility, driver/runtime support and free memory. A visible GPU does not prove that the model runtime can use it.
The complete case JSON is read successfully.
```
cat /opt/ai-lab/cases/aiops-models-runtimes-011.json
```
Common mistake to avoid: Report GPU inference merely because nvidia-smi lists a device.
2
Isolate the deciding signal
Select the runtime_backend field from the case JSON. What does it establish—and what does it not establish?
Hint & worked example
Check the runtime backend configuration; this process is still using CPU.
A successful JSON query isolates runtime_backend.
```
jq .runtime_backend /opt/ai-lab/cases/aiops-models-runtimes-011.json
```
Common mistake to avoid: Using a different case or an unrelated signal as the justification.
3
Record your operational decision
Choose the justified code: check-gpu-backend, change-without-evidence, or ignore-the-signal. Save that one code followed by a newline in /home/learner/lab/aiops-models-runtimes-011-decision.txt.
Hint & worked example
Check the runtime backend configuration; this process is still using CPU.
The decision artifact contains the correct code, not just a matching command in history.
```
printf '%s\n' 'check-gpu-backend' > /home/learner/lab/aiops-models-runtimes-011-decision.txt
```
Common mistake to avoid: Report GPU inference merely because nvidia-smi lists a device.

Understand the pattern

Accelerator availability has several layers: device visibility, driver/runtime support and free memory. A visible GPU does not prove that the model runtime can use it.

Interpret the evidence

Check the runtime backend configuration; this process is still using CPU.

A tempting mistake

Report GPU inference merely because nvidia-smi lists a device.

All names, measurements, prices, versions, and thresholds in this exercise are fictional training data. The evidence supports this bounded scenario, not a claim about a live provider.

Worked example and expected result

$ cat /opt/ai-lab/cases/aiops-models-runtimes-011.json

{
  "fixture": "Fictional training data; not a live measurement.",
  "gpu_visible": true,
  "runtime_backend": "cpu",
  "gpu_memory_free_gib": 12
}

Exit status: 0. The command completed successfully in the training fixture.

$ jq .runtime_backend /opt/ai-lab/cases/aiops-models-runtimes-011.json

"cpu"

Exit status: 0. The command completed successfully in the training fixture.

$ printf '%s\n' 'check-gpu-backend' > /home/learner/lab/aiops-models-runtimes-011-decision.txt

(no output; exit 0)

Exit status: 0. The command completed successfully in the training fixture.

Try a variation

Name the assumption that would have to change for “check-gpu-backend” to stop being the justified action. Do not silently replace missing evidence with a guess.

Explain it

What does the evidence in “Inspect accelerator availability” justify?Check the runtime backend configuration; this process is still using CPU.Report GPU inference merely because nvidia-smi lists a device.Make the change now and gather the missing evidence afterward.

Apply it elsewhere

A colleague concludes: “Report GPU inference merely because nvidia-smi lists a device.” What should you do?Accept it because the training command completed.Treat this fixture as a live measurement.Reject that conclusion; return to the case evidence and its stated limits.

Keep the momentum.

Complete the lab, then explain and apply the pattern.

Next: Spot a runtime mismatch Try an incident drill

Anonymous lesson and mastery progress lasts for this session. Sign in to keep synced lesson XP between sessions; mastery stays device-local. No speed bonus, lost lives, or hint penalties.

Keep learning

Case methodology: evidence, limits, decision