Results for “cpu”

将文案与用户自己的参考录音合成为配音、SRT 和可交接时间轴。默认使用 MiniMax 与用户自己的音色；可显式选择免费本地 Qwen3-TTS（Apple Silicon 用 MLX）或 Nano CPU。适用于口播配音、声音复刻，以及为网页视频演示准备最终音频和字幕。

The English summary is being prepared. Showing the original description.

调用 Yosys 0.35 对 RTL 进行逻辑综合并技术映射到 ASAP7 标准单元，产出门级网表 + QoR 报告。支持并行综合（基于空闲CPU数量）。Workflow: (1)生成综合脚本;(2)并行执行综合;(3)LLM检查结果并迭代优化。触发场景：(1) bba-guru-synthesis 已完成 SDC 与 CDC 检查、需要生成 netlist；(2) PD 反馈 area/path 超标需要重综合；(3) 显式 /bb-invoke-yosys。

The English summary is being prepared. Showing the original description.

Systems-level performance engineer: identifies hotspots, provides optimized code with resource justifications.

The English summary is being prepared. Showing the original description.

Forecast infrastructure and application metrics using Prophet or statsmodels for capacity planning, cost optimization, and proactive scaling. Visualize predictions in Grafana and set up alerts for projected resource exhaustion. Use when forecasting infrastructure capacity needs for CPU, memory, or disk, planning hardware procurement for next quarter, predicting cost trends to optimize cloud spending, or setting up proactive scaling policies based on predicted load.

The English summary is being prepared. Showing the original description.

Use when configuring DMA channels, circular buffer mode, double buffering, memory-to-peripheral transfers, or DMA IRQ completion on bare-metal MCUs. Not for cache theory: use cpu-cache-opt.

The English summary is being prepared. Showing the original description.

Use when diagnosing cache misses with perf, fixing false sharing, choosing AoS or SoA layout, or adding software prefetch. Not for cache theory: use memory-hierarchy-and-caches.

The English summary is being prepared. Showing the original description.

Use when explaining branch predictors, mispredict penalties, speculative execution, Spectre or Meltdown mitigations, or branchless code. Not for pipeline stage theory: use cpu-pipelines-and-hazards.

The English summary is being prepared. Showing the original description.

Use when explaining pipeline stages, data or control hazards, forwarding, stalls, or superscalar basics behind a counter reading. Not for mispredict cost: use branch-prediction-and-speculation.

The English summary is being prepared. Showing the original description.

Use when a rendering path needs stable frame-time, CPU, GPU, and memory evidence against fixed targets. Not for one-shot profiling or visual quality review.

The English summary is being prepared. Showing the original description.

Diagnose React + Three.js (or any browser canvas) frame stutter by capturing a CDP CPU profile via Playwright instead of guessing from `[perf]` console logs. Use when: (1) frame time is bad but `[perf]` instrumentation doesn't pinpoint the offender, (2) you've spent more than ~30 minutes "fixing" suspects without measuring, (3) hot paths involve animation systems, allocation bursts, or invisible scene-graph traversal, (4) you need self-time per function with line numbers, not aggregate "render slow". The `console.warn` pattern is good for monitoring; CDP is needed for diagnosis. Covers script template, sourcemap setup so minified function names are decoded, and interactive variant where the user drives the interaction and signals stop via `touch /tmp/<flag>`.

The English summary is being prepared. Showing the original description.

RFDB server becomes unresponsive (60s+ timeouts) during enricher addEdges write storms. Use when: (1) RFDB CPU spikes to 60-90% after addEdges calls; (2) subsequent reads time out at 60s; (3) server stays catatonic for 5-15 min after a write storm; (4) Tokio main thread parked (write-lock contention). Root cause: add_edges() calls maybe_auto_flush() which has a memory-pressure path that fires on NODE count — not edge count — and holds the exclusive write lock during disk I/O. Fix: separate maybe_auto_flush_edges() uses byte-limit only.

The English summary is being prepared. Showing the original description.

Analyze Ascend NPU computation-side profiling data for single-card runs or a selected rank from multi-card runs. Use this skill when the user asks to diagnose computation bottlenecks, AI Core / AI Vector / AICPU hotspots, dynamic shape overhead, block dim issues, redundant TransData/Transpose/Cast, cross-stream waits, frequency/runtime-state issues, or fusion opportunities.

The English summary is being prepared. Showing the original description.

为 msModelSlim 适配器实现逐层量化（按层加载/懒加载）能力。仅在用户明确要求逐层量化或基础适配因 CPU 内存不足无法全量加载权重时使用。该特性为高阶可选项，不是基础适配必需项。

The English summary is being prepared. Showing the original description.

Scaffold a new Tip in the ai-product-dev-tips repository following the existing directory/format conventions, and register its link in the root README.md. Use this whenever the user wants to add, create, or write a new Tip in this repo — e.g. "nlp_processing に〜の Tip を追加して", "新しい Tips を作って", "〜の解説 Tip を書いて", "this repo に〜のネタを追加". It picks the right category and next sequential number, matches the README format of nearby Tips, generates code scaffolding (Dockerfile / requirements.txt / *_cpu.sh / *_gpu.sh / *.py) only when the Tip needs runnable code, and inserts the link into the correct <details> section of the root README. Trigger this even if the user doesn't say the word "skill" — any request to add a new entry to this Tips collection should use it.

The English summary is being prepared. Showing the original description.

Use when a PlayCanvas app has too many draw calls or is CPU-bound, before hand-merging meshes or writing custom renderers.

The English summary is being prepared. Showing the original description.

Emulates position-independent shellcode in a controlled CPU emulator (Unicorn) to trace executed instructions, memory writes, and decoded second stages without running it on a real host. Activates for requests to emulate shellcode, trace a decoder stub, or recover a stage unpacked at runtime by shellcode.

The English summary is being prepared. Showing the original description.

Bypasses anti-VM and sandbox checks during analysis by locating the specific detection routines (artifact strings, timing, CPUID/hypervisor bit) and planning patches or environment hardening so the sample detonates. Activates for requests to bypass anti-VM checks, defeat sandbox detection, or make an evasive sample run for analysis.

The English summary is being prepared. Showing the original description.

Pulp in the browser — the WAM v2 and WebCLAP adapters, the wasm runtime, the Skia/WebGL2 browser window host, and the WebGPU (emdawnwebgpu) GPU-audio lane. Covers what does and does not compile to wasm, the worklet-thread constraints (no std::thread, no fetch, no navigator.gpu), the non-realtime tick, memory growth, how to PROVE a GPU lane actually ran rather than silently falling back to CPU, and the silent-failure traps that make a wasm build "work" while producing no audio or no pixels.

The English summary is being prepared. Showing the original description.

SQL discipline for querying Pulp Perfetto traces (.pftrace) with trace_processor — idempotent CREATE OR REPLACE PERFETTO views, GLOB not LIKE, dur = -1 incomplete-slice handling, EXTRACT_ARG for span args, joining on stable utid/upid, SPAN_JOIN PARTITIONED, and the draft→validate→execute loop. Ships Pulp's CPU/UI trace stdlib plus closed GPU startup, health, and probe views. TRIGGER when writing or debugging SQL over a .pftrace, when `pulp trace query` returns wrong/empty rows, or when the trace-analysis skill needs a query primitive.

The English summary is being prepared. Showing the original description.

Enable a Skia + Dawn GPU build of Pulp (MacGpuWindowHost, Skia Graphite). Covers the prebuilt skia-builder binaries, the headers-only fresh-worktree trap, reusing another checkout's cached libs via SKIA_DIR, FindSkia layouts, verifying PULP_HAS_SKIA / MacGpuWindowHost, and the raster-fallback + GPU-wedge gotchas. Use whenever GPU rendering "doesn't work" or a build silently came up CPU-only.

The English summary is being prepared. Showing the original description.

Systematic checklist for diagnosing audio artifacts. Use when the user describes a sound quality problem — clicks, pops, crackling, silence, DC offset, aliasing, phase issues, or latency problems. Trigger on phrases like "why do I hear clicks?", "my output has DC offset", "there's aliasing in my oscillator", "phase is wrong in mono", "dropout when CPU spikes". Pairs with audio-dsp-review (realtime safety) and audio-numerics-review (numeric correctness) — use those to fix once the cause is found.

The English summary is being prepared. Showing the original description.

pyfixest demeaner_backend="cupy64" (including its CPU fallback when cupy is absent) is NOT numerically identical to the default numba backend and does NOT drop fully-absorbed/collinear regressors the same way. Use when: (1) adding demeaner_backend="cupy64" to existing pf.feols/fepois calls changes the printed coefficient table, (2) a regression report suddenly gains rows with absurd estimates (e.g. coef 435.8, SE 7106) for controls absorbed by the fixed effects, (3) diffing outputs before/after a backend change, or (4) anything parses a pyfixest text report by line position.

The English summary is being prepared. Showing the original description.

Diagnose and fix slow pyfixest regression GRIDS (many feols/fepois calls run sequentially) that stay slow despite demeaner_backend="cupy64" and an idle GPU. Use when: (1) a script looping dozens of pf.feols models on a 100k+ row panel takes ~1 min/model, (2) process inspection shows ~1-1.5 cores busy and nvidia-smi shows ~0% GPU utilization with a resident cupy context, (3) planning any worker prompt that will run a model grid (robustness variants x FE structures x domains). Root cause: per-model CPU-side single-threaded fixed costs (formulaic model-matrix build, interaction construction, singleton detection, cluster vcov) dominate wall time; GPU demeaning is a small slice. Fix: shard the model grid across OS processes and/or use pyfixest multiple-estimation syntax; mandate this IN THE WORKER PROMPT.

The English summary is being prepared. Showing the original description.

Generates and updates secure, production-ready Kubernetes YAML manifests optimized for GKE Autopilot and GKE Standard clusters. Use when creating or modifying GKE deployment manifests, configuring container security contexts, setting CPU/memory resource limits, defining readiness/liveness/startup probes, mounting secrets and volumes, configuring GKE Gateway API routes, targeting Spot VMs, or deploying AI model inference workloads (vLLM, TGI, Gemma). Don't use for live cluster operations, pod troubleshooting (use gke-workload-troubleshooting), or cluster infrastructure provisioning (use gke-cluster-creation).

The English summary is being prepared. Showing the original description.