vllm-ascend-memory-profiling
Attribute Ascend HBM use during vLLM serving to weights, KV cache, HCCL, activations and runtime components.
Skill by pisceskkk
Chalk Labs
Need help deploying vllm-ascend-memory-profiling?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/pisceskkk-vllm-ascend-memory-profiling
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
- Runs as
- Python package
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/vllm-ascend-memory-profiling
What's inside
- Skill name
vllm-ascend-memory-profiling- Description
- Attribute Ascend HBM use during vLLM serving to weights, KV cache, HCCL, activations and runtime components.
- License
- MIT
- Instructions
- 22 lines in SKILL.md · view on GitHub
Before installing
- Ships executable scripts
Files (5)
- SKILL.md
- core-logic.md · references/
- msprof_fields.md · references/
- mem_analyze.py · scripts/
- weight_inspector.py · scripts/
SKILL.md (first lines)
# vLLM Ascend Memory Profiling Use this skill for the domain task described above. The workflow below is a reference contract; choose execution tools available in the target environment. ## Core workflow 1. Capture device HBM, process information, vLLM logs and msprof component data for the same run. 2. Estimate fixed overhead, weights, KV cache, HCCL and runtime with explicit evidence for each value. 3. Cross-check totals and report residual or overlapping accounting rather than silently forcing a sum. 4. State which components could not be measured and why. ## Detailed references - [msprof_fields.md](references/msprof_fields.md) ## Reference code The scripts in `scripts/` contain domain analysis or validation logic. Use them with inputs that match their documented contracts; runtime-specific tools remain external. Read [core logic](references/core-logic.md) when planning or reviewing the implementation.
From pisceskkk/vllm-ascend-skills, MIT.
More from pisceskkk/vllm-ascend-skills
- 0.586score
ascend-model-multistream-overlap-skill
pisceskkk
Design Ascend model multistream overlap from verified source mapping, resource-aware DAGs and paired validation.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
ascend-profiling-analysis
pisceskkk
Analyze Ascend profiler roots into rank, step, layer, block, operator and cross-rank findings with traceable evidence.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
ascend-triton-kernel-optimization
pisceskkk
Optimize a correctness-passed Ascend Triton kernel with per-shape measurements and profiler evidence.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
ascend-triton-kernel-validation
pisceskkk
Validate an Ascend Triton candidate against a trusted reference with a declared case matrix and fallback gate.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
ascend-triton-operator-development
pisceskkk
Develop the first correct Ascend Triton kernel from a PyTorch reference or a GPU Triton implementation.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
ascend-triton-workflow
pisceskkk
Coordinate complete Ascend Triton operator development, correctness validation and optimization with linked evidence.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
vllm-ascend-change-validation
pisceskkk
Map a vLLM or vllm-ascend diff to sufficient correctness, build and performance evidence and summarize coverage.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
vllm-ascend-correctness-validation
pisceskkk
Design and compare deterministic vLLM Ascend correctness cases across code states, execution modes and API paths.
SkillAFreshness: A★ 1Claude Code +9 - 0.584score
vllm-ascend-serving
pisceskkk
Plan and review one colocated vLLM Ascend service lifecycle, including source pairing, device selection, launch, health, and cleanup.
SkillAFreshness: A★ 1Claude Code +9 - 0.584score
vllm-ascend-distributed-debug
pisceskkk
Diagnose Ascend multi-rank startup, collective, endpoint and hang failures using explicit topology and per-rank evidence.
SkillAFreshness: A★ 1Claude Code +9 - 0.581score
vllm-ascend-pd-serving
pisceskkk
Plan and review prefill/decode vLLM Ascend serving with explicit roles, connectors, ordering and KV transfer evidence.
SkillAFreshness: A★ 1Claude Code +9 - 0.581score
vllm-ascend-performance-regression
pisceskkk
Compare baseline and candidate vLLM Ascend serving performance using controlled alternating runs.
SkillAFreshness: A★ 1Claude Code +9
Signals
- Stars
- 1
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Sep 29, 2026 (1 month ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/pisceskkk/vllm-ascend-skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from pisceskkk
- 0.586score
ascend-model-multistream-overlap-skill
pisceskkk
Design Ascend model multistream overlap from verified source mapping, resource-aware DAGs and paired validation.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
ascend-profiling-analysis
pisceskkk
Analyze Ascend profiler roots into rank, step, layer, block, operator and cross-rank findings with traceable evidence.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
ascend-triton-kernel-optimization
pisceskkk
Optimize a correctness-passed Ascend Triton kernel with per-shape measurements and profiler evidence.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
ascend-triton-kernel-validation
pisceskkk
Validate an Ascend Triton candidate against a trusted reference with a declared case matrix and fallback gate.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
ascend-triton-operator-development
pisceskkk
Develop the first correct Ascend Triton kernel from a PyTorch reference or a GPU Triton implementation.
SkillAFreshness: A★ 1Claude Code +9 - 0.586score
ascend-triton-workflow
pisceskkk
Coordinate complete Ascend Triton operator development, correctness validation and optimization with linked evidence.
SkillAFreshness: A★ 1Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get vllm-ascend-memory-profiling running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
vllm-ascend-memory-profiling skill. Agent Mart. https://agentmart.dev/skills/pisceskkk-vllm-ascend-memory-profiling (last updated 2026-10-24)