AgentSkillsCN

parallel-computing

在可量化的扩展证据支持下，设计、优化并验证跨CPU线程/工作者的并行执行。在选择并行化策略、诊断竞争与负载失衡、评估加速/效率曲线、调优任务粒度，或对基线与当前并行性能的回归问题进行分类时使用此技能。

SKILL.md
--- frontmatter
name: parallel-computing
description: Design, optimize, and validate parallel execution across CPU threads/workers with measurable scaling evidence. Use when selecting parallelization strategy, diagnosing contention and load imbalance, evaluating speedup/efficiency curves, tuning task granularity, or triaging baseline-vs-current parallel performance regressions.

Parallel Computing

Use this skill to convert parallel performance work into reproducible scaling decisions.

Workflow

  1. •Define scaling objective and constraints.
  • •Capture workload shape, data size, and latency/throughput targets.
  • •Define hardware assumptions (core count, SMT policy, NUMA context).
  1. •Choose parallel model and partitioning.
  • •Select task/data/pipeline parallelism intentionally.
  • •Set chunk size and scheduling strategy to minimize overhead and imbalance.
  • •Define shared-state boundaries before coding.
  1. •Diagnose bottlenecks.
  • •Check lock contention, false sharing, synchronization frequency, and memory bandwidth pressure.
  • •Separate algorithmic limits from runtime/scheduler overhead.
  1. •Validate scaling behavior.
  • •Compare baseline vs current throughput by thread count.
  • •Evaluate parallel efficiency and regressions at each thread level.
  • •Treat regressions above threshold as blockers.
  1. •Deliver implementation handoff.
  • •Include tuning deltas, tradeoffs, and reproducible benchmark commands.
  • •Provide clear patch plan for runtime/algorithm changes.

Commands

bash
python3 scripts/compare_parallel_scaling.py \
  --baseline <baseline.json> \
  --current <current.json> \
  --regression-threshold-pct 5 \
  --efficiency-drop-threshold-pct 10

Treat non-zero exits as blocker regressions.

Output Contract

Return:

  1. •Scaling Context: workload and hardware assumptions.
  2. •Findings: thread-level throughput/speedup/efficiency deltas.
  3. •Optimization Plan: concrete runtime/algorithm changes.
  4. •Verification: benchmark commands and thresholds.
  5. •Residual Risks: unresolved contention or scaling ceilings.

References

  • •references/workflow.md: detailed parallel optimization sequence.
  • •references/scaling-playbook.md: common bottlenecks and remedies.
  • •references/signoff-template.md: concise scaling sign-off format.

Execution Rules

  • •Compare like-for-like workloads and environments only.
  • •Report both speedup and efficiency, not throughput alone.
  • •Flag thread-level regressions above thresholds as blockers.
  • •Avoid overfitting to one thread count; evaluate full scaling curve.