Skip to content

Rocm subversion enhancement #123

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Open
jaspals3123 wants to merge 5 commits into
base: development
Choose a base branch
from
Open

Rocm subversion enhancement #123

jaspals3123 wants to merge 5 commits into from

Conversation

@jaspals3123
Copy link
Collaborator

@jaspals3123 jaspals3123 commented Feb 6, 2026

Run:
node-scraper run-plugins RocmPlugin --exp-rocm 7.0.1-42 --exp-rocm-sub-versions '{"version-lrt":"7.0.1-42", "version-rocm-developer-tools": "7.0.1-42"}'

jaspals3123 and others added 3 commits February 4, 2026 15:23
@jaspals3123 @cursoragent
Add ROCm sub-version validation with result tracking
7cb7819 
- Add check_exp_rocm_sub_versions method that returns bool
- Track validation results using boolean flags before setting final status
- Update result.message to include all validation outcomes
- Prevent result override by collecting all check results first
- Follow pattern similar to package_analyzer and dkms_analyzer

Co-authored-by: Cursor <cursoragent@cursor.com>
@jaspals3123
initial changes
73d01fc
@jaspals3123
adding tests
bd06fc2
alexandraBara
alexandraBara requested changes Feb 9, 2026
View reviewed changes
test/functional/fixtures/rocm_plugin_config.json
Comment on lines +6 to +8
"exp_rocm": "7.0.0-38",
"exp_rocm_sub_versions": {
"version_rocm": "7.0.0-38"
Copy link
Collaborator

@alexandraBara alexandraBara Feb 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

wouldnt the "subversion" part be whatever is after the ":". exp_rocm arg should stay as it is, the exp_rocm_sub_versions should check the subversion part. Here it looks like subversion is just another way to say rocm_version, i dont see the colon

Copy link
Collaborator Author

@jaspals3123 jaspals3123 Feb 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

as per ES implementation everthing under rocm/.info/* is considered subversion where one of the line is version_rocm but we can replace expected value under exp_rocm_subverdion with anything from below
r$ grep . -r /opt/rocm/.info/* /opt/rocm/.info/version:7.0.1 /opt/rocm/.info/version-hiprt:7.0.1-42 /opt/rocm/.info/version-hiprt-devel:7.0.1-42 /opt/rocm/.info/version-lrt:7.0.1-42 /opt/rocm/.info/version-oclrt:7.0.1-42 /opt/rocm/.info/version-ocl-sdk:7.0.1-42 /opt/rocm/.info/version-openmp:7.0.1-42 /opt/rocm/.info/version-rocm:7.0.1-42 /opt/rocm/.info/version-rocm-developer-tools:7.0.1-42 /opt/rocm/.info/version-rocm-hip:7.0.1-42

nodescraper/cli/cli.py
Comment on lines +341 to +344
# Handle comma-separated plugin names (but not option values like JSON)
arg_stripped = arg.strip()
looks_like_value = arg_stripped.startswith("{") or arg_stripped.startswith("[")
if not arg.startswith("-") and "," in arg and not looks_like_value:
Copy link
Collaborator

@alexandraBara alexandraBara Feb 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

we have not made the decision to allow comma separated plugins, so lets remove this

test/unit/plugin/test_rocm_analyzer.py
Comment on lines +55 to +57
"version": "6.2.0-66",
"version-rocm": "6.2.0-66",
"version-hip-sdk": "6.2.0-66",
Copy link
Collaborator

@alexandraBara alexandraBara Feb 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

same here, what is the difference between rocm_version="6.2.0-66" and rocm_sub_version? where version="6.2.0-66"

Copy link
Collaborator Author

@jaspals3123 jaspals3123 Feb 9, 2026 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

here version and version-rocm is output of :

`(venv) (py39) jaspals@ppac-cyxtera-cx77-5:~/node-scraper$ grep . -r /opt/rocm/.info/version-rocm

7.0.1-42

(venv) (py39) jaspals@ppac-cyxtera-cx77-5:~/node-scraper$ grep . -r /opt/rocm/.info/version

7.0.1
`

subversions is anything present under this cmd like version-hiprt:7.0.1-42, also version-lrt:7.0.1-42 etc :

(venv) (py39) jaspals@ppac-cyxtera-cx77-5:~/node-scraper$ grep . -r /opt/rocm/.info/* /opt/rocm/.info/version:7.0.1 /opt/rocm/.info/version-hiprt:7.0.1-42 /opt/rocm/.info/version-hiprt-devel:7.0.1-42 /opt/rocm/.info/version-lrt:7.0.1-42 /opt/rocm/.info/version-oclrt:7.0.1-42 /opt/rocm/.info/version-ocl-sdk:7.0.1-42 /opt/rocm/.info/version-openmp:7.0.1-42 /opt/rocm/.info/version-rocm:7.0.1-42 /opt/rocm/.info/version-rocm-developer-tools:7.0.1-42 /opt/rocm/.info/version-rocm-hip:7.0.1-42

@alexandraBara
Copy link
Collaborator

alexandraBara commented Feb 10, 2026

I think we are adding too many error messages somewhere, the count for errors is 2 but its only displaying one. Either we are not displaying both errors or we there should just be 1 error count:

(venv) [alexbara@ppac-cyxtera-cr60-4 node-scraper]$ node-scraper --plugin-config ~/plugin_config_rocm.json
  2026-02-10 14:49:21 UTC       INFO               nodescraper | Log path: ./scraper_logs_ppac_cyxtera_cr60_4_ctr_dcgpu_2026_02_10-02_49_21_PM
  2026-02-10 14:49:21 UTC       INFO               nodescraper | System Name: ppac-cyxtera-cr60-4.ctr.dcgpu
  2026-02-10 14:49:21 UTC       INFO               nodescraper | System SKU: None
  2026-02-10 14:49:21 UTC       INFO               nodescraper | System Platform: None
  2026-02-10 14:49:21 UTC       INFO               nodescraper | System location: SystemLocation.LOCAL
  2026-02-10 14:49:21 UTC       INFO               nodescraper | Initializing connection manager for InBandConnectionManager with default args
  2026-02-10 14:49:21 UTC       INFO               nodescraper | --------------------------------------------------
  2026-02-10 14:49:21 UTC       INFO               nodescraper | Running plugin RocmPlugin
  2026-02-10 14:49:21 UTC       INFO               nodescraper | Initializing connection: InBandConnectionManager
  2026-02-10 14:49:21 UTC       INFO               nodescraper | Using local shell
  2026-02-10 14:49:21 UTC       INFO               nodescraper | Checking OS family
  2026-02-10 14:49:21 UTC       INFO               nodescraper | OS Family: LINUX
  2026-02-10 14:49:21 UTC       INFO               nodescraper | Running data collector: RocmCollector
  2026-02-10 14:49:21 UTC       INFO               nodescraper | (RocmPlugin) ROCm version: 6.4.2-120
  2026-02-10 14:49:21 UTC       INFO               nodescraper | Running data analyzer: RocmAnalyzer
  2026-02-10 14:49:21 UTC   CRITICAL               nodescraper | ROCm version mismatch! Expected: ['7.0.0-38'], actual: 6.4.2-120
  2026-02-10 14:49:21 UTC      ERROR               nodescraper | ROCm sub-version mismatch!
  2026-02-10 14:49:21 UTC      ERROR               nodescraper | (RocmPlugin) ROCm version mismatch!. ROCm sub-version mismatch (2 errors: ROCm version mismatch! Expected: ['7.0.0-38'], actual: 6.4.2-120, ROCm sub-version mismatch!)
  2026-02-10 14:49:21 UTC       INFO               nodescraper | Closing connections
  2026-02-10 14:49:21 UTC       INFO               nodescraper | Running result collators
  2026-02-10 14:49:21 UTC       INFO               nodescraper | Running TableSummary result collator
  2026-02-10 14:49:21 UTC       INFO               nodescraper |

+-------------------------+--------+-----------------------------+
| Connection              | Status | Message                     |
+-------------------------+--------+-----------------------------+
| InBandConnectionManager | OK     | task completed successfully |
+-------------------------+--------+-----------------------------+

+------------+--------+------------------------------------------------------------------------------+
| Plugin     | Status | Message                                                                      |
+------------+--------+------------------------------------------------------------------------------+
| RocmPlugin | ERROR  | Analysis error: ROCm version mismatch!. ROCm sub-version mismatch (2 errors: |
|            |        | ROCm version mismatch! Expected: ['7.0.0-38'], actual: 6.4.2-120, ROCm sub-  |
|            |        | version mismatch!)                                                           |
+------------+--------+------------------------------------------------------------------------------+

  2026-02-10 14:49:21 UTC       INFO               nodescraper | Data written to csv file: ./scraper_logs_ppac_cyxtera_cr60_4_ctr_dcgpu_2026_02_10-02_49_21_PM/nodescraper.csv

Im using the plugin config from the functional test:

{
  "global_args": {},
  "plugins": {
    "RocmPlugin": {
      "analysis_args": {
        "exp_rocm": "7.0.0-38",
        "exp_rocm_sub_versions": {
          "version_rocm": "7.0.0-38"
        }
      }
    }
  },
  "result_collators": {},
  "name": "RocmPlugin config",
  "desc": "Config for testing RocmPlugin"
}

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@alexandraBara alexandraBara alexandraBara requested changes

@landrews-amd landrews-amd Awaiting requested review from landrews-amd landrews-amd is a code owner

Requested changes must be addressed to merge this pull request.

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@jaspals3123 @alexandraBara