Tools - Add runner for sys info and update docs

docs/cli.md

@@ -165,6 +165,26 @@ Execute GPT2 model benchmark in default configuration:
 sb exec --config-override superbench.enable="['gpt2_models']"
 ```
 
+### `sb node info`
+Get system info on the local node.
+
+```bash title="SB CLI"
+sb node info [--output-dir]
+```
+
+#### Optional arguments
+
+| Name           | Default | Description                                                                 |
+|----------------|---------|-----------------------------------------------------------------------------|
+| `--output-dir` | `None`  | Path to output directory, outputs/{datetime} will be used if not specified. |
+
+#### Examples
+
+Get system info on the local node and save it into the `outputs` dir:
+```bash title="SB CLI"
+sb node info --output-dir outputs
+```
+
 ### `sb result diagnosis`
 
 Filter the defective machines automatically from benchmarking results according to rules defined in rule file.
@@ -284,6 +304,7 @@ sb run [--config-file]
        [--docker-image]
        [--docker-password]
        [--docker-username]
+       [--get-info]
        [--host-file]
        [--host-list]
        [--host-password]
@@ -302,6 +323,7 @@ sb run [--config-file]
 | `--docker-image` `-i`    | `superbench/superbench` | Docker image URI.                                                           |
 | `--docker-password`      | `None`                  | Docker registry password if authentication is needed.                       |
 | `--docker-username`      | `None`                  | Docker registry username if authentication is needed.                       |
+| `--get-info`             | `False`                 | Collect system info.                                                        |
 | `--host-file` `-f`       | `None`                  | Path to Ansible inventory host file.                                        |
 | `--host-list` `-l`       | `None`                  | Comma separated host list.                                                  |
 | `--host-password`        | `None`                  | Host password or key passphase if needed.                                   |
@@ -335,6 +357,16 @@ sb run --no-docker --host-list localhost --config-override \
   superbench.enable=kernel-launch superbench.env.SB_MICRO_PATH=/path/to/superbenchmark
 ```
 
+Collect system info on all nodes in ./host.ini" distributed without running benchmarks:
+```bash title="SB CLI"
+sb run --get-info --host-file ./host.ini -C superbench.enable=none
+```
+
+Collect system info on all nodes in ./host.ini" distributed while running benchmarks:
+```bash title="SB CLI"
+sb run --get-info --host-file ./host.ini
+```
+
 ### `sb version`
 
 Print the current SuperBench CLI version.

docs/user-tutorial/system-config.md

@@ -4,6 +4,8 @@ id: system-config
 
 # System Config Info
 
+This tool is to collect the system information automatically on the tested GPU nodes including the following hardware categories:
+
 - [System](#system)
 - [Memory](#memory)
 - [CPU](#cpu)
@@ -12,7 +14,33 @@ id: system-config
 - [Accelerator](#accelerator)
 - [PCIe](#pcie)
 
-## Parameter amd Details
+## Usage
+
+### Usage on local machine
+
+1. [Install SuperBench](../getting-started/installation.mdx) on the local machine using root privilege.
+
+2. Start to collect the sys info using `sb node info --output-dir ${output-dir}` command using root privilege.
+
+3. After the command finished, you can find the output system info json file `sys-info.json` of local node under \${output_dir}.
+
+### Usage on multiple remote machines
+
+1. [Install SuperBench](../getting-started/installation.mdx) on the local machine.
+
+2. [Deploy SuperBench](../getting-started/run-superbench.md#deploy) onto the remote machines.
+
+2. Prepare the host file of the tested GPU nodes using [Ansible Inventory](../getting-started/configuration.md#ansible-inventory) on the local machine.
+
+3. After installing the Superbnech and the host file is ready, you can start to collect the sys info automatically using  `sb run --get-info` command. The detailed command can be found from [SuperBench CLI](../cli.md).
+
+  ```
+  sb run --get-info -f host.ini --output-dir ${output-dir} -C superbench.enable=none
+  ```
+
+4. After the command finished, you can find the output system info json file `sys-info.json` of each node under \${output_dir}/nodes/${node_name}.
+
+## Parameter and Details
 
 ### System
 

superbench/cli/_commands.py

@@ -67,6 +67,9 @@ def load_arguments(self, command):
                 nargs='+',
                 help='Extra arguments to override config_file.'
             )
+            ac.argument(
+                'get_info', options_list=('--get-info', '-g'), action='store_true', help='Collect node system info.'
+            )
 
         with ArgumentsContext(self, 'benchmark') as ac:
             ac.argument('name', options_list=('--name', '-n'), type=str, help='Benchmark name or regular expression.')

superbench/cli/_handler.py

@@ -275,7 +275,8 @@ def run_command_handler(
     output_dir=None,
     private_key=None,
     config_file=None,
-    config_override=None
+    config_override=None,
+    get_info=False,
 ):
     """Run the SuperBench benchmarks distributedly.
 
@@ -295,6 +296,7 @@ def run_command_handler(
         config_file (str, optional): Path to SuperBench config file. Defaults to None.
         config_override (str, optional): Extra arguments to override config_file,
             following [Hydra syntax](https://hydra.cc/docs/advanced/override_grammar/basic). Defaults to None.
+        get_info (bool, optional): Collect node system info. Defaults to False.
 
     Raises:
         CLIError: If input arguments are invalid.
@@ -316,6 +318,10 @@ def run_command_handler(
     )
 
     runner = SuperBenchRunner(sb_config, docker_config, ansible_config, sb_output_dir)
+
     runner.run()
+    if get_info:
+        runner.run_sys_info()
+
     if runner.get_failure_count() != 0:
         sys.exit(runner.get_failure_count())

superbench/cli/_help.py

@@ -63,6 +63,10 @@
           text: >
             {cli_name} run --no-docker --host-list localhost
             --config-override superbench.enable=kernel-launch superbench.env.SB_MICRO_PATH=/path/to/superbenchmark
+        - name: Collect system info on all nodes in ./host.ini" without running benchmarks
+          text: {cli_name} run --get-info --host-file ./host.ini -C superbench.enable=none
+        - name: Collect system info on all nodes in ./host.ini" while running benchmarks
+          text: {cli_name} run --get-info --host-file ./host.ini
 """.format(cli_name=CLI_NAME)
 
 helps['benchmark'] = """

superbench/cli/_node_handler.py

@@ -3,17 +3,28 @@
 
 """SuperBench CLI node subgroup command handler."""
 
+from pathlib import Path
+import json
+
 from superbench.tools import SystemInfo
+from superbench.common.utils import create_sb_output_dir
 
 
-def info_command_handler():
+def info_command_handler(output_dir=None):
     """Get node hardware info.
 
+    Args:
+        output_dir (str): Output directory.
+
     Returns:
         dict: node info.
     """
     try:
         info = SystemInfo().get_all()
+        output_dir = create_sb_output_dir(output_dir)
+        output_dir_path = Path(output_dir)
+        with open(output_dir_path / 'sys_info.json', 'w') as f:
+            json.dump(info, f)
     except Exception as ex:
         raise RuntimeError('Failed to get node info.') from ex
     return info

superbench/runner/runner.py

@@ -199,6 +199,24 @@ def deploy(self):    # pragma: no cover
             )
         self._ansible_client.run(self._ansible_client.get_playbook_config('deploy.yaml', extravars=extravars))
 
+    def run_sys_info(self):
+        """Run the system info on all nodes."""
+        self.check_env()
+
+        logger.info('Runner is going to get node system info.')
+
+        fcmd = "docker exec sb-workspace bash -c '{command}'"
+        if self._docker_config.skip:
+            fcmd = "bash -c 'cd $SB_WORKSPACE && {command}'"
+        ansible_runner_config = self._ansible_client.get_shell_config(
+            fcmd.format(command='sb node info --output-dir {output_dir}'.format(output_dir=self._sb_output_dir))
+        )
+        ansible_rc = self._ansible_client.run(ansible_runner_config, sudo=(not self._docker_config.skip))
+
+        if ansible_rc != 0:
+            self.cleanup()
+        self.fetch_results()
+
     def check_env(self):    # pragma: no cover
         """Check SuperBench environment."""
         logger.info('Checking SuperBench environment.')