Understand experiment configuration

Define where to run and what to execute in two TOML files. Start with an inline workload; separate reusable definitions later.

Two files

File	Responsibility
System	Scheduler, resources, paths, and shared settings.
Scenario	Test cases, workload settings, and execution requirements.

The Run your first experiment uses standalone.toml and an inline Fio scenario:

name = "quickstart-fio"

[[Tests]]
id = "fio"
name = "fio"
description = "Small file I/O benchmark"
test_template_name = "Fio"
num_nodes = 1
time_limit = "00:05:00"

  [Tests.cmd_args]
  fio_binary = "fio"

    [Tests.cmd_args.args]
    name = "cloudai-quickstart"
    filename = "/tmp/cloudai-fio-quickstart"
    rw = "randwrite"
    bs = "128k"
    size = "80m"
    iodepth = 1
    numjobs = 1
    group_reporting = true
    thread = true

test_template_name selects the integration. cmd_args configures its workload. id identifies the case within the scenario; it is distinct from the scenario’s name.

Reuse a definition

When several scenarios share a workload definition, move its name, template, and arguments into a test TOML. Use test_name in each case and pass the definition directory with --tests-dir. Keep allocation settings such as num_nodes and time_limit in the scenario.

Add complexity deliberately

Use Compose a scenario for multiple cases and dependencies, Prepare and clean up with hooks for preparation and cleanup, and Run a configuration search for configuration search. List-valued settings can become search dimensions; see Agent and search settings before using lists as ordinary workload arguments. Exact system fields are in Systems; case fields are in Scenario configuration.