platform-dev

Develop HyperParallel platform abstraction: new Platform APIs, torch/mindspore backends, FSDP/HSDP/Pipeline/Activation Checkpoint, DTensorBase, collectives. Use when changing platform/ or cross-backend features. Not for shard op YAML/impl (use dist-op-dev) or git/CI (autogit / gate-doctor).

Jump to install

Repository: mindspore-ai/hyper-parallel
Last source activity: August 7, 2026 at 03:17
Detected SKILL.md language: English
Stars: 7
Forks: 4

Install options

The review-first prompt is selected by default. You can switch to a direct command or download a local copy.

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

A direct command skips the review prompt. Inspect the source before running it.

npx skills add https://github.com/mindspore-ai/hyper-parallel --skill platform-dev

The command stays on one line. Scroll horizontally to inspect it before copying.

Prefer a local copy? Download the files currently available to SkillsMP.

Review the source files

Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.

File Explorer

10 files

SKILL.md

Source instructions · Read-only preview

Back to search Back to installation

add-unit-test

Add or extend HyperParallel unit tests under tests/ut. Use when writing UT, increasing coverage, or mocking distributed/hardware. Not for multi-card ST (see testing.md / distributed-op-testing) or shard-op-specific UT constraints (distributed-op-testing rule).

2026-08-07

autogit

GitCode fork workflow: commit, push, create/append PR, status, squash, lint/test. Triggers: 帮我提交, create PR, PR 状态, /commit, /create-pr, etc. Origin=fork, upstream=main. Details in references/.

2026-08-07

code-review

Review HyperParallel changes for distributed correctness, stream sync, memory safety, cross-platform parity, and code quality. Use for /code-review, PR review, or when the user asks to review/check changes.

2026-08-07

View all from this repository

name	platform-dev
description	Develop HyperParallel platform abstraction: new Platform APIs, torch/mindspore backends, FSDP/HSDP/Pipeline/Activation Checkpoint, DTensorBase, collectives. Use when changing platform/ or cross-backend features. Not for shard op YAML/impl (use dist-op-dev) or git/CI (autogit / gate-doctor).

Platform Development

Guides cross-platform work under platform/. This file is the index — load the matching workflow / reference before that step.

When to use

New method on platform/platform.py, or impl in platform/torch|mindspore/
FSDP / HSDP / Pipeline / Activation Checkpoint platform code
DTensorBase extensions; collectives; stream sync / memory lifecycle patterns

Execution (autonomous checklist)

Tree, API map, patterns: references/architecture.md, references/quick-reference.md, references/decisions.md.

Hard rules

No bare import torch / mindspore in platform-agnostic code — get_platform()
New Platform APIs land in base class first (platform/platform.py)
Consider both backends — implement or NotImplementedError
Type gaps: torch device/ProcessGroup vs mindspore str groups — check signatures
Lazy framework imports inside methods under platform/torch|mindspore/ + # pylint: disable=C0415
Stream/memory: handle.wait(), event.record→wait, resize_(0) — see .agent/rules/distributed.md

Out of scope

Distributed op YAML / parallel_*.py → dist-op-dev
Commit / PR / CI → autogit / gate-doctor
Full distributed audit after change → code-review

platform-dev

Source facts

Install options

Review the source files

More from this repository

More from this repository

Platform Development

When to use

Execution (autonomous checklist)

Hard rules

Out of scope

Platform Development

When to use

Execution (autonomous checklist)

Hard rules

Out of scope