fix: allow unsetting of tasks for gpu jobs

[cmeesters]

this PR should fix #316 - python's process management seems to interfere with SLURM tasks under unknown circumstances. Unsetting the tasks per gpu might be a solution.

If a user defined tasks <= 0, the --ntasks-per-gpu flag is not issued.

Summary by CodeRabbit

• New Features

  • Added support for a new resource key to control the number of tasks per GPU in SLURM job submissions.

• Bug Fixes

  • Improved handling of SLURM --ntasks-per-gpu and --ntasks options, allowing greater flexibility and correct omission when appropriate.

[coderabbitai]

Walkthrough

The update changes how SLURM submission commands are constructed for jobs requesting GPUs. It introduces a tasks_per_gpu resource to control the --ntasks-per-gpu option, allowing users to override or omit this flag. For non-GPU jobs, the --ntasks flag is now always set. Associated comments are updated and repositioned.

Changes

File(s)	Change Summary
snakemake_executor_plugin_slurm/submit_string.py	Refactored logic for setting --ntasks-per-gpu and --ntasks in SLURM command; added tasks_per_gpu resource; updated comments.

Sequence Diagram(s)

sequenceDiagram
    participant User
    participant SnakemakeJob
    participant SubmitString

    User->>SnakemakeJob: Define job resources (gpus, tasks, tasks_per_gpu)
    SnakemakeJob->>SubmitString: Call get_submit_command(job, params)
    alt Job requests GPUs
        alt tasks_per_gpu >= 1
            SubmitString->>SubmitString: Add --ntasks-per-gpu <tasks_per_gpu>
        else tasks_per_gpu < 1
            SubmitString->>SubmitString: Omit --ntasks-per-gpu
        end
    else Non-GPU job
        SubmitString->>SubmitString: Add --ntasks <tasks>
    end
    SubmitString-->>SnakemakeJob: Return SLURM command string

Loading

Possibly related PRs

• fix: add lost code back in #254: Addresses GPU-related task allocation logic in SLURM job submission, focusing on similar resource flags but in a different module.

Suggested reviewers

• johanneskoester

Poem

A tweak for the tasks and the GPUs,
Now jobs can choose what flags to use.
With tasks_per_gpu you set the pace,
No more defaults you can’t erase!
Rabbits rejoice, the logic is neat—
SLURM commands now hop to your beat! 🐇

---

📜 Recent review details

Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 6d7c50a and 7f1a817.

📒 Files selected for processing (1)

• snakemake_executor_plugin_slurm/submit_string.py (1 hunks)

🧰 Additional context used

🧠 Learnings (2)

📓 Common learnings

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#0
File: :0-0
Timestamp: 2025-01-13T09:54:22.950Z
Learning: PR #173 (adding gres resource specification) depends on PR #28 in snakemake-executor-plugin-slurm-jobstep repository, as changes were required in the cpu-settings function of the jobstep-Executor module.

Learnt from: johanneskoester
PR: snakemake/snakemake-executor-plugin-slurm#173
File: docs/further.md:96-96
Timestamp: 2025-03-10T15:20:51.829Z
Learning: PR #173 in snakemake-executor-plugin-slurm implements GPU job support by adding resources: `gres` for generic resource specifications (e.g., 'gpu:1'), `gpu`/`gpus` for specifying GPU counts, and `gpu_model`/`gpu_manufacturer` for specifying GPU types, allowing users to request GPU resources directly rather than having to use slurm_extra.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#187
File: .github/workflows/post_to_mastodon.yml:0-0
Timestamp: 2025-01-17T17:27:32.446Z
Learning: In the mastodon publishing workflow for snakemake-executor-plugin-slurm, the PR_TITLE environment variable is required by the post_to_mastodon.sh script and should be preserved.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#173
File: tests/tests.py:79-86
Timestamp: 2025-03-10T13:34:59.270Z
Learning: In the Snakemake executor plugin for SLURM, all GPU resources (both simple numeric values and model:number specifications) should be translated to the `--gpus` flag (plural form) to match SLURM's command line interface expectations.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#173
File: tests/tests.py:79-86
Timestamp: 2025-03-10T13:34:59.270Z
Learning: In the Snakemake executor plugin for SLURM, all GPU resources (even simple numeric ones) should be translated to the `--gpus` flag (plural) to match SLURM's expectations, not `--gpu` (singular).

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#193
File: .github/workflows/post_to_mastodon.yml:26-26
Timestamp: 2025-01-20T09:13:26.443Z
Learning: In the snakemake-executor-plugin-slurm repository, release PRs follow the naming pattern "chore(main): release X.Y.Z" where X.Y.Z is the version number.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#221
File: snakemake_executor_plugin_slurm/efficiency_report.py:0-0
Timestamp: 2025-06-16T08:54:07.957Z
Learning: In the context of SLURM executor plugin development, user cmeesters (experienced with SLURM) noted they are not aware of SLURM typically reporting memory values with decimals (like 1.5G), but accepted defensive programming to handle such cases anyway.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#173
File: snakemake_executor_plugin_slurm/__init__.py:0-0
Timestamp: 2025-02-18T14:40:27.064Z
Learning: In the Snakemake executor plugin for SLURM, the GPU resource is specified using the "gpus" (plural) resource name, not "gpu" (singular).

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#173
File: snakemake_executor_plugin_slurm/utils.py:75-100
Timestamp: 2025-02-18T14:49:42.624Z
Learning: In the Snakemake SLURM executor plugin, users must specify either `gres` or `gpus` for GPU resources, but not both simultaneously, as these are mutually exclusive ways to request GPU resources.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#173
File: tests/tests.py:79-86
Timestamp: 2025-03-10T13:22:54.081Z
Learning: In the Snakemake executor plugin for SLURM, when using GPU resources, the function returns `--gpu=n` (singular) when only a number is provided, but `--gpus=model:n` (plural) when a model is specified in the implementation.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#249
File: snakemake_executor_plugin_slurm/submit_string.py:21-29
Timestamp: 2025-03-31T11:52:05.373Z
Learning: In the snakemake-executor-plugin-slurm project, account and partition parameters are validated in the submit function before reaching the get_submit_command function. By design, these parameters are either empty strings or already properly formatted with SLURM parameter tags, making additional validation in get_submit_command unnecessary.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#221
File: snakemake_executor_plugin_slurm/efficiency_report.py:34-49
Timestamp: 2025-05-26T12:22:07.005Z
Learning: In the parse_reqmem function in snakemake_executor_plugin_slurm/efficiency_report.py, the conditional check for per_unit == "/node" is necessary for proper conversion of memory values in multi-node jobs, not redundant as it might initially appear.

snakemake_executor_plugin_slurm/submit_string.py (10)

Learnt from: johanneskoester
PR: snakemake/snakemake-executor-plugin-slurm#173
File: docs/further.md:96-96
Timestamp: 2025-03-10T15:20:51.829Z
Learning: PR #173 in snakemake-executor-plugin-slurm implements GPU job support by adding resources: `gres` for generic resource specifications (e.g., 'gpu:1'), `gpu`/`gpus` for specifying GPU counts, and `gpu_model`/`gpu_manufacturer` for specifying GPU types, allowing users to request GPU resources directly rather than having to use slurm_extra.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#173
File: tests/tests.py:79-86
Timestamp: 2025-03-10T13:34:59.270Z
Learning: In the Snakemake executor plugin for SLURM, all GPU resources (both simple numeric values and model:number specifications) should be translated to the `--gpus` flag (plural form) to match SLURM's command line interface expectations.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#173
File: tests/tests.py:79-86
Timestamp: 2025-03-10T13:34:59.270Z
Learning: In the Snakemake executor plugin for SLURM, all GPU resources (even simple numeric ones) should be translated to the `--gpus` flag (plural) to match SLURM's expectations, not `--gpu` (singular).

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#173
File: tests/tests.py:79-86
Timestamp: 2025-03-10T13:22:54.081Z
Learning: In the Snakemake executor plugin for SLURM, when using GPU resources, the function returns `--gpu=n` (singular) when only a number is provided, but `--gpus=model:n` (plural) when a model is specified in the implementation.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#173
File: snakemake_executor_plugin_slurm/__init__.py:0-0
Timestamp: 2025-02-18T14:40:27.064Z
Learning: In the Snakemake executor plugin for SLURM, the GPU resource is specified using the "gpus" (plural) resource name, not "gpu" (singular).

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#173
File: snakemake_executor_plugin_slurm/utils.py:75-100
Timestamp: 2025-02-18T14:49:42.624Z
Learning: In the Snakemake SLURM executor plugin, users must specify either `gres` or `gpus` for GPU resources, but not both simultaneously, as these are mutually exclusive ways to request GPU resources.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#0
File: :0-0
Timestamp: 2025-01-13T09:54:22.950Z
Learning: PR #173 (adding gres resource specification) depends on PR #28 in snakemake-executor-plugin-slurm-jobstep repository, as changes were required in the cpu-settings function of the jobstep-Executor module.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#249
File: snakemake_executor_plugin_slurm/submit_string.py:21-29
Timestamp: 2025-03-31T11:52:05.373Z
Learning: In the snakemake-executor-plugin-slurm project, account and partition parameters are validated in the submit function before reaching the get_submit_command function. By design, these parameters are either empty strings or already properly formatted with SLURM parameter tags, making additional validation in get_submit_command unnecessary.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#178
File: snakemake_executor_plugin_slurm/__init__.py:0-0
Timestamp: 2024-12-11T14:17:08.749Z
Learning: In the `snakemake-executor-plugin-slurm` project, when handling exceptions in `snakemake_executor_plugin_slurm/__init__.py`, prefer concise error messages and avoid unnecessary verbosity or exception chaining when it's not necessary.

Learnt from: cmeesters
PR: snakemake/snakemake-executor-plugin-slurm#221
File: snakemake_executor_plugin_slurm/efficiency_report.py:34-49
Timestamp: 2025-05-26T12:22:07.005Z
Learning: In the parse_reqmem function in snakemake_executor_plugin_slurm/efficiency_report.py, the conditional check for per_unit == "/node" is necessary for proper conversion of memory values in multi-node jobs, not redundant as it might initially appear.

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)

• GitHub Check: testing

🔇 Additional comments (2)

snakemake_executor_plugin_slurm/submit_string.py (2)

55-62: LGTM! Solid implementation for allowing users to disable --ntasks-per-gpu.

The new tasks_per_gpu resource parameter with conditional flag setting effectively addresses issue #316. The fallback to the existing tasks parameter maintains backward compatibility while allowing users to disable the problematic flag by setting tasks_per_gpu <= 0.

---

64-66: Good refactoring - cleaner separation of GPU vs non-GPU job logic.

Moving the --ntasks setting for non-GPU jobs into the else branch improves code organization and makes the distinction between GPU and non-GPU job handling clearer.

✨ Finishing Touches

• 📝 Generate Docstrings

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

🪧 Tips

Chat

There are 3 ways to chat with CodeRabbit:

• Review comments: Directly reply to a review comment made by CodeRabbit. Example:
  • I pushed a fix in commit <commit_id>, please review it.
  • Explain this complex logic.
  • Open a follow-up GitHub issue for this discussion.
• Files and specific lines of code (under the "Files changed" tab): Tag @coderabbitai in a new review comment at the desired location with your query. Examples:
  • @coderabbitai explain this code block.
  • @coderabbitai modularize this function.
• PR comments: Tag @coderabbitai in a new PR comment to ask questions about the PR branch. For the best results, please provide a very specific query, as very limited context is provided in this mode. Examples:
  • @coderabbitai gather interesting stats about this repository and render them as a table. Additionally, render a pie chart showing the language distribution in the codebase.
  • @coderabbitai read src/utils.ts and explain its main purpose.
  • @coderabbitai read the files in the src/scheduler package and generate a class diagram using mermaid and a README in the markdown format.
  • @coderabbitai help me debug CodeRabbit configuration file.

Support

Need help? Create a ticket on our support page for assistance with any issues or questions.

Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments.

CodeRabbit Commands (Invoked using PR comments)

• @coderabbitai pause to pause the reviews on a PR.
• @coderabbitai resume to resume the paused reviews.
• @coderabbitai review to trigger an incremental review. This is useful when automatic reviews are disabled for the repository.
• @coderabbitai full review to do a full review from scratch and review all the files again.
• @coderabbitai summary to regenerate the summary of the PR.
• @coderabbitai generate docstrings to generate docstrings for this PR.
• @coderabbitai generate sequence diagram to generate a sequence diagram of the changes in this PR.
• @coderabbitai resolve resolve all the CodeRabbit review comments.
• @coderabbitai configuration to show the current CodeRabbit configuration for the repository.
• @coderabbitai help to get help.

Other keywords and placeholders

• Add @coderabbitai ignore anywhere in the PR description to prevent this PR from being reviewed.
• Add @coderabbitai summary to generate the high-level summary at a specific location in the PR description.
• Add @coderabbitai anywhere in the PR title to generate the title automatically.

Documentation and Community

• Visit our Documentation for detailed information on how to use CodeRabbit.
• Join our Discord Community to get help, request features, and share feedback.
• Follow us on X/Twitter for updates and announcements.

[cmeesters]

This is a draft PR for now - requires documentation and more testing.