Pipeline GPU Kernel

Apply software pipelining (double-buffering) to a tiled GPU kernel to overlap global memory loads with Tensor Core computation. Covers prologue/loop/epilogue restructuring, LDG-register vs cp.async (LDGSTS) variant selection based on compute/load ratio, shared memory budget verification against architecture-specific occupancy cliffs, and SASS-level verification of load/compute overlap.

45 stars 4 forksProductivity & Notes

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add pjt222/agent-almanac --skill pipeline-gpu-kernel
Manual install
git clone --depth 1 https://github.com/pjt222/agent-almanac.git /tmp/agent-almanac
mkdir -p ~/.claude/skills
cp -r /tmp/agent-almanac/i18n/caveman-lite/skills/pipeline-gpu-kernel ~/.claude/skills/pipeline-gpu-kernel

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Pipeline GPU Kernel

Apply software pipelining (double-buffering) to a tiled GPU kernel so that global memory loads for tile N+1 overlap with Tensor Core computation on tile N. Transform a sequential load-sync-compute-sync K-loop into a prologue/loop/epilogue structure, choose between LDG-register and cp.async (LDGSTS) variants based on compute/load ratio, verify shared memory stays under the architecture occupancy cliff, and confirm load/compute overlap in the final SASS.

## When to Use

- When `analyze-kernel-bottleneck` identifies a memory-bound kernel with low compute/load ratio per tile
- When warp interleaving alone cannot hide DRAM latency (~300 cycles on GA104)
- When the kernel has a sequential load-sync-compute-sync K-loop that can be restructured
- Not needed when compute/load ratio is high (>20:1) and 8+ warps are active

## Inputs

- **Required**: CUDA kernel source file (`.cu`) with a tiled K-loop containing separate load and compute phases
- **Required**: Target GPU architecture (e.g., GA104 / sm_86 — determines smem cliff and occupancy limits)
- **Required**: Current tile sizes (BM, BN, BK) and data type (FP16, FP32, INT8)
- **Optional**: Compute/load ratio per tile (from `analyze-kernel-bottleneck`; will be estimated if not provided)
- **Optional**: Benchmark baseline (non-pipelined performance at target problem size)

## Procedure

### Step 1: Verify Preconditions
Read the full skill on GitHub

Details

Skill name
pipeline-gpu-kernel
Path
i18n/caveman-lite/skills/pipeline-gpu-kernel/SKILL.md
License
MIT
Repo updated
18 November 2026

FAQ

What does the Pipeline GPU Kernel skill do?

Apply software pipelining (double-buffering) to a tiled GPU kernel to overlap global memory loads with Tensor Core computation. Covers prologue/loop/epilogue restructuring, LDG-register vs cp.async (LDGSTS) variant selection based on compute/load ratio, shared memory budget verification against architecture-specific occupancy cliffs, and SASS-level verification of load/compute overlap.

How do I install the Pipeline GPU Kernel skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add pjt222/agent-almanac --skill pipeline-gpu-kernel. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Pipeline GPU Kernel popular?

It lives in pjt222/agent-almanac, which has 45 GitHub stars and 4 forks, and is included in 1 curated skill list.

Related skills

Adapt Architecture

45

pjt222/agent-almanac

Execute structural metamorphosis using strangler fig migration, chrysalis phases, and interface preservation. Covers transformation planning, parallel running, progressive cutover, rollback design, and post-metamorphosis stabilization for system architecture evolution. Use when assess-form has classified the system as READY for transformation, when migrating from monolith to microservices, when replacing a core subsystem while dependents continue operating, or when any architectural change must be gradual rather than big-bang.

Software DevelopmentIn 1 list

Adaptic

45

pjt222/agent-almanac

Master skill composing the 5-step synoptic cycle for panoramic synthesis across multiple domains. Orchestrates meditate, expand-awareness, observe, awareness, integrate-gestalt, and express-insight into a coherent process that produces unified understanding rather than sequential compromise. Use when a problem genuinely spans 3+ domains and the interactions between domains matter more than depth in any one, when sequential analysis feels like compromise rather than integration, or before major architectural decisions affecting multiple stakeholders.

OtherIn 1 list

Add Puzzle Type

45

pjt222/agent-almanac

Scaffold a new puzzle type across all 10+ pipeline integration points in jigsawR. Creates the core puzzle module, wires it into the unified pipeline (generation, positioning, rendering, adjacency), adds ggpuzzle geom/stat layers, updates DESCRIPTION and config.yml, extends the Shiny app, and creates a comprehensive test suite. Use when adding a completely new puzzle type to the package or following the 10-point integration checklist to ensure nothing is missed end-to-end.

Testing & QAIn 1 list

Plan Capacity

45

pjt222/agent-almanac

Perform capacity planning using historical metrics and growth models. Use predict_linear for forecasting, identify resource constraints, calculate headroom, and recommend scaling actions before saturation. Use before seasonal traffic spikes or product launches, during quarterly capacity reviews, when resource utilization trends upward, or before budget planning cycles.

Productivity & NotesIn 1 list

Oc Status

43

AlsKozlov/ru-legal

Weekly status emails к outside counsel — per-matter skeletons → markdown output + email drafts. Используй когда пользователь говорит "weekly OC update", "запросить статус у advocate", "что нового от внешних юристов".

Productivity & Notes

Plan Garden Calendar

45

pjt222/agent-almanac

Plan garden activities using solar, lunar, and biodynamic calendars. Covers USDA hardiness zones, frost date calculation, equinox/solstice anchoring, synodic lunar cycle (waxing/waning), ascending/descending moon, Maria Thun biodynamic calendar (root/leaf/flower/fruit days), succession planting schedules, and seasonal task planning. Use when planning a new growing season and needing a planting schedule, integrating lunar or biodynamic timing into garden practice, calculating frost dates and planting windows for a specific zone, setting up succession planting for continuous harvest, or conducti

Productivity & NotesIn 1 list

Browse the most popular Claude skills, all productivity & notes skills, or write your own with the skills generator.